edgar mauricio produÇÃo de biodiesel usando …

Universidade de Aveiro

2021

EDGAR MAURICIO VARGAS SOLANO

PRODUÇÃO DE BIODIESEL USANDO MATERIAIS RESIDUAIS BIODIESEL PRODUCTION USING RESIDUAL MATERIALS

Universidade de Aveiro

2021

EDGAR MAURICIO VARGAS SOLANO

PRODUÇÃO DE BIODIESEL USANDO MATERIAIS RESIDUAIS BIODIESEL PRODUCTION USING RESIDUAL MATERIALS

Tese apresentada à Universidade de Aveiro para cumprimento dos requisitos necessários à obtenção do grau de Doutor em Engenharia Química, realizada sob a orientação científica da Doutora Maria Isabel da Silva Nunes, Professora Auxiliar do Departamento de Ambiente e Ordenamento da Universidade de Aveiro

Financial support from: Direction of Investigation, Creation and Extension of Jorge Tadeo Lozano University, FCT/MCTES for the financial support to

CESAM (UIDP/50017/2020 & UIDB/50017/2020) and FCT, I.P. for the research contract CEECIND/00383/2017 under the CEEC Individual 2017.

I dedicate this doctoral thesis to my parents (Anita and Eliseo) who are in heaven and who have always been an example for my life.

o júri

Presidente Professor Doutor João Manuel Nunes Torrão Professor Catedrático, Universidade de Aveiro

Professor Doutor Abel Gomes Martins Ferreira

Professor Auxiliar, Universidade de Coimbra

Doutora Ana Sofia Ramos Brásio

Responsável pelo Desenvolvimento de Processos da Prio Energy

Professor Doutor Luís António da Cruz Tarelho

Professor Associado, Universidade de Aveiro

Professora Doutora Nídia de Sá Caetano

Professora Coordenadora, Instituto Superior de Engenharia do Porto

Professora Doutora Maria Isabel da Silva Nunes

Professora Auxiliar, Universidade de Aveiro (orientadora)

acknowledgements

The execution of this study only became reality thanks to people and organizations I have to thank: First, to the professor Dra. Maria Isabel da Silva Nunes, who gave me a great learning experience, a great person and an exceptional professional. To professor Doctor Luís António Tarelho (Associate Professor of the Department of Environment and Planning of the University of Aveiro), to professor Doctor João Manuel Coutinho (Full Professor at the Chemistry Department of University of Aveiro) and Dra. Márcia Carvalho Neves (PhD), auxiliary researcher (CICECO and Department of Chemistry, University of Aveiro). To the Direction of Research, Creation and Extension of the Jorge Tadeo Lozano University, CESAM, CICECO, Department of Environment and Planning and Department of Chemistry. I thank my family, who were present at every moment and especially my wife Adriana, my son Daniel Felipe and my daughter Maria Paula. To my friends, both the ones I have (Alis, Yineth, Jorge, Andres, Ingrid, Sebastian, Laura, Duvan, Lizeth, Isaac ...) and the ones I did during my doctorate (Joaquim-Kim, Daniel, Ivonne), you are part of this work. Finally, I thank everyone who encouraged and supported me throughout this journey.

palavras-chave

Materiais residuais, FAME, catalisador sólido, catalisadores bifuncionais, óleo de cozinha residual, óleo de palma refinado, otimização de processos.

resumo

A energia é um requisito básico para a existência humana e seu consumo aumenta a cada ano. Atualmente, a maior parte das necessidades energéticas é suprida pelos convencionais recursos de origem fóssil, tais como a gasolina, o gás liquefeito de petróleo, o petrodiesel e o gás natural. No entanto, o uso de combustíveis fósseis tem associados sérios problemas ambientais. Uma das fontes de energia alternativa (renovável) mais promissora é o biodiesel. Normalmente a produção de biodiesel é um processo catalisado, no qual são utilizados catalisadores alcalinos ou ácidos, para a conversão de triglicerídeos (reação de transesterificação) e ácidos gordos livres (reação de esterificação) em ésteres metílicos de ácidos gordos (FAME), utilizando metanol. Neste contexto, recentemente os catalisadores heterogéneos têm chamado a atenção dos investigadores pelas suas vantagens relativamente aos catalisadores homogéneos, nomeadamente em termos de maior rendimento de biodiesel, maior pureza do glicerol, separação mais fácil do catalisador (filtração simples), menos corrosivos e ambientalmente “mais amigáveis”. O presente trabalho teve como objetivo preparar um catalisador sólido bifuncional eficiente (i.e., capaz de catalisar simultaneamente as reações de transesterificação e esterificação), a partir de materiais residuais, com vista à produção de FAME a partir de misturas de óleos vegetais de baixo custo (óleo alimentar usado (OAU) e óleo de palma refinado (OPR)), e metanol. Também constituiu um objetivo de trabalho a otimização do processo de produção tanto num reator descontínuo como num reator de leito fixo contínuo. A cinza volante de biomassa (FAD) foi selecionada, de entre outros materiais residuais (rocha dolomítica natural, casca de ovo de galinha e tereftalato de polietileno - PET) estudados, por ter exibido melhor desempenho na produção de FAME, e também deter um caráter bifuncional. O rendimento máximo de FAME alcançado (processo não otimizado), na etapa de seleção dos materiais, foi cerca de 96 % (m/m), a 60 oC, 9:1 (mol/mol) razão metanol/óleo, 10 % (m/m) de carga de FAD, durante 180 min em reator descontínuo. Na etapa de otimização da produção em reator descontínuo, usando FAD, o rendimento máximo de FAME registado foi de 73.8 % para as condições operacionais: 13.6 % (m/m) de carga de catalisador, 6.7 de razão molar metanol / óleo, 72 % (m/m) de OAU/OPR e 55 oC. Observou-se que catalisador pode ser usado pelo menos até três ciclos consecutivos sem perda de atividade catalítica. Na otimização da produção de FAME em reator de leito fixo contínuo, com a FAD peletizada, a concentração máxima de FAME registada foi de 89.7 %, nas seguintes condições operacionais: 124 min de tempo de residência, 74.6 % (m/m) de OAU/OPR, 12:1 razão molar metanol/óleo e 60 °C. O catalisador manteve-se estável ao longo de 32 h de operação contínua, sem desativação perceptível. O presente trabalho dá um contributo para tornar o processo de produção de biodiesel mais alinhado com os princípios da economia circular, através da integração de resíduos, quer com uma função catalítica (FAD), quer como matéria-prima (OAU). Selecionou-se um material residual (cinza volante de biomassa) que necessita de um tratamento físico muito simples (secagem), para ser utilizado num processo descontínuo de produção de FAME. Além disso, este material na forma peletizada, continua a exibir excelentes propriedades catalíticas, permitindo assim a produção de FAME em contínuo. Os resultados deste trabalho são muito promissores, sobretudo quando se perspetiva uma futura aplicação industrial.

keywords

Waste materials, FAME, solid catalyst, bifunctional catalysts, waste cooking oil, refined palm oil, process optimization.

abstract

Energy is a basic requirement for human existence and its consumption increases every year. Currently, most of the energy needs come from conventional fossil fuels, such as gasoline, liquefied petroleum gas, petrodiesel and natural gas. However, the use of fossil fuels has serious environmental problems associated with it. One of the most promising alternative (renewable) energy sources is biodiesel. Usually the production of biodiesel is a catalyzed process, in which alkaline or acid catalysts are used, for the conversion of triglycerides (transesterification reaction) and free fatty acids (esterification reaction) into fatty acid methyl esters (FAME), using methanol. In this context, heterogeneous catalysts have recently attracted the attention of researchers due to their advantages over the homogeneous ones in terms of higher biodiesel yield, higher glycerol purity, easier catalyst separation (simple filtration), less corrosive and more environmentally "friendly”. The present work aimed to prepare an efficient bifunctional solid catalyst (capable of simultaneously catalyzing the transesterification and esterification reactions), from residual materials, to produce FAME from low-cost vegetable oil mixtures (waste cooking oil (WCO) and refined palm oil (RPO)) and methanol. The optimization of the production process in both a batch reactor and a continuous fixed-bed reactor is also an objective of the work. The biomass fly ash (FAD) was selected from among other studied residual raw materials (natural dolomite rock, chicken eggshells and polyethylene terephthalate - PET), for exhibiting the best performance in the production of FAME, and also for having a bifunctional character. The maximum yield of FAME reached (non-optimized process), in the raw materials selection stage, was about 96 wt%, at 60 oC, 9:1 (mol/mol) methanol/oil ratio, 10 wt% FAD loading, for 180 min in a batch reactor. In the production optimization stage in a batch reactor, using FAD, the maximum FAME yield reached was 73.8 % for the operating conditions: 13.57 wt% of catalyst loading, 6.7 of methanol/oil molar ratio, 72 wt% of WCO/RPO at 55 oC. It was observed that the catalyst could be used in up to three consecutive cycles without loss of catalytic activity. For the optimization in the continuous fixed bed reactor, with the pelletized FAD, the maximum concentration of FAME reached was 89.7 %, under the following operating conditions: 124 min residence time, 74.6 wt% of WCO/RPO, 12:1 methanol/oil molar ratio at 60 oC. The catalyst was stable during the 32 hours of continuous operation, without significant deactivation. The present work contributes to the biodiesel production process being more aligned with the principles stablished by circular economy, by integrating waste, either with a catalytic function (FAD) or as a raw material (WCO). A residual material that needs a very simple physical treatment (drying) was selected to be used (biomass fly ash) in a discontinuous process for the production of FAME. Furthermore, this material in granulated form continues to exhibit excellent catalytic properties, thus allowing continuous production of FAME. The results of this work are very promising, especially when looking towards a future industrial application.

Palabras clave

Materiales de desecho, FAME, catalizador sólido, catalizadores bifuncionales, aceite de cocina usado, aceite de palma refinado, optimización de procesos.

resumen

La energía es un requisito básico para la existencia humana y su consumo aumenta cada año. Actualmente, la mayor parte de las necesidades energéticas proviene de combustibles fósiles convencionales, como la gasolina, el gas licuado del petróleo, el petrodiesel y el gas natural. Sin embargo, el uso de combustibles fósiles tiene asociados serios problemas ambientales. Una de las fuentes de energía alternativa (renovable) más prometedoras es el biodiesel. Normalmente la producción de biodiesel es un proceso catalizado, en el cual son utilizados catalizadores alcalinos o ácidos, para la conversión de triglicéridos (reacción de transesterificación) y ácidos grasos libres (reacción de esterificación) en ésteres metílicos de ácidos grasos (FAME), utilizando metanol. En este contexto, recientemente los catalizadores heterogéneos han llamado la atención de los investigadores por sus ventajas frente a los homogéneos en términos de mayor rendimiento de biodiesel, mayor pureza de glicerol, separación más fácil del catalizador (filtración simple), menos corrosivos y ambientalmente “más amigables”. El presente trabajo tuvo como objetivo preparar un catalizador sólido bifuncional eficiente (es decir, capaz de catalizar simultaneamente las reacciones de transesterificación y esterificación), a partir de materiales residuales, para producir FAME a partir de mezclas de aceites vegetales de bajo costo (aceite de cocina residual (ACU) y aceite de palma refinado (APR)) y metanol. También constituye un objetivo del trabajo la optimización del proceso de producción tanto en un reactor discontinuo como en un reactor continuo de lecho fijo. Las cenizas volantes de biomasa (FAD) fueron seleccionadas de entre otras materias primas residuales (roca de dolomita natural, cáscaras de huevo de gallina y tereftalato de polietileno - PET) estudiadas, por haber exhibido el mejor desempeño en la producción de FAME, y también por tener un caracter bifuncional. El rendimiento máximo de FAME alcanzado (proceso no optimizado), en la etapa de selección de las materias primas fue cerca del 96 % (m/m), a 60 oC, 9:1 (mol/mol) relación metanol/aceite, 10 % (m/m) de carga de FAD, durante 180 min en reactor discontinuo. En la etapa de optimización de la producción en un reactor discontinuo, usando FAD, el rendimiento máximo de FAME alcanzado fue del 73.8 % para las condiciones de operación: 13.57 % (m/m) de carga de catalizador, 6.7 de relación molar metanol/aceite, 72 % (m/m) de WCO/RPO a 55 oC. Se observó que el catalizador pudo ser usado hasta tres ciclos consecutivos sin pérdida de actividad catalítica. Para la optimización en el reactor de lecho fijo continuo, con el FAD peletizado, la concentración máxima de FAME alcanzada fue del 89.7 %, en las siguientes condiciones de operación: 124 min de tiempo de residencia, 74.6 % (m/m) de WCO/RPO, 12:1 relación molar metanol/aceite y 60 oC. El catalizador se mantuvo estable durante las 32 h de funcionamiento continuo, sin una desactivación apreciable. El presente trabajo contribuye a que el proceso de producción de biodiésel esté más alineado con los principios de la economía circular, mediante la integración de residuos, ya sea con función catalítica (FAD) o como materia prima (ACU). Se seleccionó un material residual (cenizas volantes de biomasa) que necesita un tratamiento físico muy simple (secado), para ser utilizado en un proceso discontinuo de producción de FAME. Además, este material en forma granulada, continúa exhibiendo excelentes propiedades catalíticas, permitiendo así la producción de FAME en continuo. Los resultados de este trabajo son muy prometedores, especialmente cuando se mira hacia una futura aplicación industrial.

TABLE OF CONTENTS

xvii

Table of contents

Table of contents ..................................................................................................................... xvii

List of figures ............................................................................................................................ xxi

List of tables .............................................................................................................................xxiii

Notation ................................................................................................................................... xxv

SECTION A - The study proposal ......................................................................................... 1

1 Presentation of the thesis .................................................................................................... 3

1.1 Introduction .................................................................................................................. 3

1.2 Thesis objectives and structure .................................................................................. 10

1.3 Scientific production within the scope of the thesis .................................................. 13

1.3.1 Articles published in scientific journals ............................................................... 13

1.3.2 Articles published in scientific conferences ........................................................ 13

1.3.3 Participation in scientific events .......................................................................... 14

References ............................................................................................................................. 16

SECTION B – Bifunctional solid catalyst selection .............................................................. 19

2 Solid catalysts obtained from wastes for FAME production using mixtures of refined

palm oil and waste cooking oils ................................................................................................. 21

2.1 Introduction ................................................................................................................ 22

2.2 Materials and Methods ............................................................................................... 23

2.2.1 Materials .............................................................................................................. 24

2.2.2 Oil mixtures (RPO and WCO) characterization .................................................... 24

2.2.3 Catalysts preparation and characterization ........................................................ 25

2.2.4 FAME synthesis .................................................................................................... 26

2.3 Results and discussion ................................................................................................ 28

2.3.1 Oil mixtures characterization .............................................................................. 28

2.3.2 Catalysts characterization ................................................................................... 29

2.3.2.1 BET surface area and Hammett indicators analyses ....................................... 29

2.3.2.2 SEM and EDX analyses ..................................................................................... 31

BIODIESEL PRODUCTION FORM RESIDUAL MATERIALS

xviii

2.3.2.3 XRD analyses .................................................................................................... 34

2.3.2.4 FTIR analyses .................................................................................................... 37

2.3.2.5 Catalysts performance ..................................................................................... 40

2.4 Conclusions ................................................................................................................. 45

References ............................................................................................................................. 47

SECTION C - Optimization of FAME production in batch mode operation .......................... 53

3 Optimization of FAME production from blends of waste cooking oil and refined palm oil

using biomass fly ash as a catalyst ............................................................................................ 55

3.1 Introduction ................................................................................................................ 56

3.2 Materials and Methods .............................................................................................. 57

3.2.1 Materials.............................................................................................................. 58



3.2.4 FAME synthesis and quantification ..................................................................... 60

3.2.5 Experimental design and optimization of FAME production process................. 61

3.2.6 Catalyst reusability .............................................................................................. 63



3.3.2 Catalysts characterization ................................................................................... 64

3.3.2.1 BET surface area and Hammett indicators analyses ....................................... 65

3.3.2.2 SEM and EDX analyses ..................................................................................... 66

3.3.2.3 XRD analyses .................................................................................................... 67

3.3.2.4 FTIR analyses .................................................................................................... 68

3.3.3 Optimization of FAME production process: regression model and statistical

analysis 68

3.3.4 Catalyst reusability: catalytic performance assessment ..................................... 76

3.4 Conclusions ................................................................................................................. 79

References ............................................................................................................................. 81

SECTION D - Optimization of FAME production in continuous fixed-bed reactor ............... 85

4 Pelletized biomass fly ash for FAME production: optimization of a continuous process . 87

4.1 Introduction ................................................................................................................ 88

4.2 Materials and Methods .............................................................................................. 90

TABLE OF CONTENTS

xix



4.2.3 Experimental setup .............................................................................................. 92

4.2.4 FAME content ...................................................................................................... 94

4.2.5 Preliminary assays for steady state identification .............................................. 95

4.2.6 Experimental design for optimization of the FAME production ......................... 97



4.3.2 Catalysts characterization ................................................................................. 100

4.3.2.1 XRD analysis ................................................................................................... 100

4.3.2.2 Hammett indicators and BET surface area analyses ..................................... 101

4.3.2.3 SEM analysis ................................................................................................... 102

4.3.2.4 FTIR analysis ................................................................................................... 102

4.3.2.5 XPS analysis .................................................................................................... 103

4.3.3 Optimization of FAME production process: regression model and statistical

analysis 104

4.3.4 Catalytic stability of the pelletized biomass ash ............................................... 112

4.4 Conclusions ............................................................................................................... 113

References ........................................................................................................................... 115

SECTION E – Final remarks .............................................................................................. 121

5 Final remarks .................................................................................................................... 123

5.1 General conclusions .................................................................................................. 123

5.2 Future works ............................................................................................................. 126

Annexes .................................................................................................................................... A1


xx

LIST OF FIGURES

xxi

List of figures

Figure 1.1 - Simplified process flow chart of alkali-catalyzed biodiesel production (adapted

from Leung et al. (2010)). ......................................................................................... 4

Figure 1.2 - Global transesterification reaction........................................................................... 5

Figure 1.3 - Esterification reaction. ............................................................................................. 5

Figure 1.4 - Two-step conventional catalysis versus bifunctional catalyst process. ................... 7

Figure 1.5 - Structure of the thesis in five sections. .................................................................. 15

Figure 2.1 - FAD catalyst: SEM (a) and EDX (b and c); FAC catalyst: SEM (e) and EDX (d and f).

................................................................................................................................ 32

Figure 2.2 - Dolomite C: SEM (a); Dolomite CSC: SEM (b) and EDX (c). .................................... 33

Figure 2.3 - CaO-SiO2: SEM (a): CaO-S-SiO2: EDX (b) and (c). .................................................... 33

Figure 2.4 - SEM images of CA-PET (a) and CA-PET-S (b) catalysts. .......................................... 34

Figure 2.5 - XRD patterns of catalysts: FAD and FAC (a), Dolomite C and Dolomite CSC (b),

CaO-SiO2 and CaO-S-SiO2 (c), and CA-PET and CA-PET-S (d). (• SiO2, ■ CaCO3, ♦

CaO, ▲ CaSO4, ■ CaCO3, ▼ Ca(OH)2, ♦ CaO, □ Ca2SiO4 and ★ MgO). .................. 36

Figure 2.6 - FTIR spectra of catalysts: FAD and FAC (a), Dolomite C and Dolomite CSC (b), CaO-

SiO2 and CaO-S-SiO2 (c), and CA-PET and CA-PET-S (d). ......................................... 38

Figure 2.7 - Performance in terms of FAME yield and FFA conversion of catalysts: FAD and

FAC (a) and (b), Dolomite C and Dolomite CSC (c) and (d), CaO-SiO2 and CaO-S-

SiO2 (e) and (f), and CA-PET and CA-PET-S(g) and (h), for several RPO:WCO

mixtures. ................................................................................................................. 42

Figure 3.1 - Adsorption and desorption isotherms for the FAD catalyst. ................................. 65

Figure 3.2 - FAD catalyst characterization by: SEM (a and b) and EDX (c). ............................... 66

Figure 3.3 - XRD patterns of FAD catalyst.................................................................................. 67

Figure 3.4 - FTIR spectrum of FAD catalyst. ............................................................................... 68

Figure 3.5 - (a) Residual normal probability plot, (b) Residual versus predicted response plot,

(c) Predicted versus actual values plot. .................................................................. 73


xxii

Figure 3.6 - Response surface plots of FAME yield as a function of: (a) RPO/WCO ratio and

catalyst loading at 50 oC and methanol/oil = 6 mol/mol; (b) RPO/WCO ratio and

methanol/oil at 50 oC and catalyst loading = 10 wt%; (c) RPO/WCO ratio and

temperature for catalyst loading = 10 wt% and methanol/oil = 6 mol/mol; (d)

temperature and methanol/oil ratio for catalyst loading = 10 wt% and RPO/WCO

= 50 wt%; (e) temperature (oC) and catalyst loading (wt%) for methanol/oil = 6

mol/mol and RPO/WCO = 50 wt%; (f) methanol/oil ratio and catalyst loading

(wt%) at 50 oC and RPO/WCO = 50 wt%. ............................................................... 74

Figure 3.7 - Reusability studies of the FAD catalyst under the optimal operating conditions. 77

Figure 3.8 - XRD patterns of FAD catalyst for the different reuse cycles. ................................. 78

Figure 4.1- Biomass fly ash catalyst: (a) powder (before pelletization), (b) pelletized. ........... 91

Figure 4.2 - Experimental setup diagram used in this work (1. Feed zone, 2. Reaction zone, 3.

Separation zone). ................................................................................................... 93

Figure 4.3 - Assembly of the fixed bed reaction system for continuous FAME production. .... 94

Figure 4.4 - Preliminary assays to identify the steady state of the continuous fixed bed

reactor. ................................................................................................................... 97

Figure 4.5 - XRD Diffractogram of biomass fly ash catalyst (• SiO2, ■ CaCO3, ▲ KCl, ★ Ca

(OH)2). ................................................................................................................... 100

Figure 4.6 - Absorption and desorption isotherms of biomass fly ash. .................................. 101

Figure 4.7 - SEM images of biomass fly ash using a magnification of 500X (a) and 5000X (b).

.............................................................................................................................. 102

Figure 4.8 - FTIR spectrum of biomass fly ash catalyst. .......................................................... 103

Figure 4.9 - The XPS wide energy spectrum of biomass fly ash catalyst. ............................... 104


(c) Predicted versus experimental values plot. .................................................... 109

Figure 4.11 - FAME concentration response surface graphics based on: (a) Residence time

and WCO/RPO ratio (wt%) at molar methanol/oil ratio of 9 mol/mol; (b) Time of

residence and molar methanol/oil ratio at WCO/RPO ratio of 50 wt%; (c) Molar

methanol/oil ratio and WCO/RPO ratio (wt%) at time of residence of 120 min. 110

Figure 4.12 - FAME concentration of the catalyst over 32 h of continuous operation. ......... 113

LIST OF TABLES

xxiii

List of tables

Table 2.1 - Solid catalysts preparation procedures. .................................................................. 25

Table 2.2 - Properties of the oil mixtures used. ........................................................................ 28

Table 2.3 - Textural properties of the catalysts prepared in this work. .................................... 29

Table 3.1 - Range and factor levels of operating variables used in the Box – Behnken

experimental design. .............................................................................................. 62

Table 3.2 - Properties of the oil mixtures used in this work. .................................................... 64

Table 3.3 - Experimental design and predicted results of RSM. ............................................... 69

Table 3.4 - ANOVA results of the response surface quadratic model without interactions. .... 70

Table 3.5 - ANOVA results for the coefficients of the variables in the quadratic regression

model without interactions. ................................................................................... 71

Table 3.6 - Textural properties and acid/basic strength of the FAD catalyst used in three

FAME synthesis cycles. ........................................................................................... 78

Table 4.1 - Operating conditions of continuous flow reactors to produce FAME. ................... 96

Table 4.2 - Factor and levels of process variables used in the Box – Behnken experimental

design. ..................................................................................................................... 98

Table 4.3 - Properties of the oil mixtures used. ........................................................................ 99

Table 4.4 - Summary of XPS analysis done on the surface of biomass fly ash catalyst. ......... 104

Table 4.5 - Experimental and predicted results of RSM. ......................................................... 105

Table 4.6 - ANOVA table of the regression model. ................................................................. 106

Table 4.7 - Regression coefficients (coded factors) for the fitting quadratic polynomial model.

.............................................................................................................................. 107

Table 4.8 - Experimental data on optimal operating conditions............................................. 112


xxiv

NOMENCLATURE

xxv

Notation

List of abbreviations

ASTM American Society for Testing and Materials

B10 10 % biodiesel and 90 % diesel

B20 20 % biodiesel and 80 % diesel

BET Brunauer, Emmett y Teller

CA Activated Carbon

CA-PET Activated Carbon of Polyethylene Terephthalate

CA-PET-S Activated Carbon of Polyethylene Terephthalate Sulfonation

CESAM Center for Environmental and Marine Studies

CFBR Continuous Fixed Bed Reactor

Dolomite C Calcined Dolomite (catalyst)

Dolomite CSC Calcined Sulfonated Calcined Dolomite

EDX Energy Dispersive X-ray Spectroscopy

FAC Biomass fly ash (calcined)

FAD Biomass fly ash (dried)

FAME Fatty Acid Methyl Esters

FFA Free Fatty Acids

FTIR Fourier Transform Infrared

ICONTEC Colombian Institute of Technical Standards and Certification

NTC Colombian technical standard


xxvi

PAH Polycyclic Aromatic Hydrocarbons

PET Polyethylene Terephthalate

RPO Refined Palm Oil

RSM Response Surface Methodology

RT Residence Time

SEM Scanning Electron Microscope

SN Saponification Number

WCO Waste Cooking Oils

XPS X-ray photoelectron spectroscopy

XRD X-ray diffraction

Nomenclature

Roman Letters

𝐴 total peak areas of the methyl ester

𝐴𝐸𝐼 the peak area of methyl heptadecanoate

𝐴𝑉 acid value (mg KOH/g)

𝐴𝑉𝑖 acid value of the initial oil mixture (mg KOH/g)

𝐴𝑉𝑓 acid value of the final oil mixture (mg KOH/g)

𝛽𝑖 linear terms of polynomial equation

𝛽𝑖𝑖 quadratic terms of polynomial equation

NOMENCLATURE

xxvii

𝛽𝑖𝑗 interaction terms of polynomial equation

𝛽0 intercept coefficient (offset)

𝐶 concentration of FAME

C.V. coefficient of variation

M mixtures of WCO and RPO

MW molecular weight (g/mol)

pKa acidity measure

pKb basicity measure

R coefficient statistics of fit

S.D. standard deviation

𝑇 temperature (K) or (oC)

𝑊 the mass of the sample for analysis (mg)

𝑊𝐸𝐼 the mass of methyl heptadecanoate (mg)

𝑥𝑖 independent variable

𝑥𝑗 independent variable

𝑌 predicted response


xxviii

Greek letters

𝜀 measurement error

λ wavelength (Å)

θ angle of incidence

1

SECTION A - The study proposal

In Section A the subject of the study is introduced, which includes the motivation and the

relevance of the work carried out. The general and specific objectives are also established

as well as the thesis structure. The purpose of Section A is to show the reader the

importance of biodiesel production from waste materials which is fully aligned with the

principles of circular economy, and how the work is organized to achieve the proposed

objectives.


2

THE STUDY PROPOSAL

3

1 Presentation of the thesis

1.1 Introduction

Energy is a basic requirement for human existence and its consumption increases all years. In

the current situation, the foremost amount of primary and useful energy is supplied by the

conventional fossil fuel resources, such as gasoline, liquefied petroleum gas, diesel fuel, and

natural gas. However, the use of fossil fuels has several influences on the environment, such

as large amount of greenhouse gas emissions, acid rain, resources depletion, etc. In addition

to serious environmental issues, dwindling reserves of crude oil, fluctuating petroleum fuel

prices, have made today’s need to find alternative “green” sources of energy, which are

sustainable, environmentally compatible, economically competitive, and easily available.

One of the most promising sources is biodiesel, an alternative diesel fuel derived from

renewable sources with high quality, which allows the replacement of fossil diesel oil (Leung

et al. 2010). Its advantages over petroleum diesel cannot be overemphasized: it is safe,

renewable, non-toxic, biodegradable, sulfur free, and good lubricating properties. In addition,

its use engenders numerous societal benefits, namely rural revitalization, creation of new jobs,

and reduced global warming.

Biodiesel has significant influences in reducing engine emissions such as unburned

hydrocarbons (68 %), particulate matter (40 %), carbon monoxide (44 %), sulfur oxide

(100 %), and polycyclic aromatic hydrocarbons (PAH) (80–90 %) (Kiss et al. 2008).

Most of the biodiesel is produced by the alkali-catalyzed process. Figure 1.1 shows a simplified

flow diagram of the alkaline catalyst process. Feedstock with high free fatty acids will

undesirably react with the alkaline catalyst to form soap. The maximum amount of free fatty

acids acceptable in an alkali catalyzed system is less than 2.5 wt%. If the oil or fat feedstock

BIODIESEL PRODUCTION USING RESIDUAL MATERIALS

4

has a FFA content greater than 2.5 wt%, a pretreatment step is necessary before the

transesterification process.

Figure 1.1 - Simplified process flow chart of alkali-catalyzed biodiesel production (adapted

from Leung et al. (2010)).

Usually the biodiesel production is a catalyzed process where alkali or acid compounds are

used for the conversion of triglycerides and free fatty acids (FFA) into fatty acid methyl esters

(FAME), when methanol is used in the synthesis, as shown in Figure 1.2 and Figure 1.3,

respectively. The use of homogeneous alkaline catalyst such as NaOH and KOH for the

transesterification of waste cooking oil to biodiesel posed serious separation problem. This is

mainly due to the high FFA content in waste cooking oil that reacts, in the presence of water,

with the alkaline catalyst to form soap. The formation of soap will cause serious difficulty in

the separation of biodiesel and the glycerin (a reaction product) and, consequently the amount

of water needed for washing the FAME will be higher, resulting in decreased yield to FAME; all

THE STUDY PROPOSAL

5

of these incur in high production costs and high volume of wastewater to be treated (Lam et

al. 2009). Another alternative method (implemented in two-steps) is to use homogeneous acid

catalyst such as H2SO4 and HCl to initially reduce the FFA content through esterification

reaction (first stage) and only then followed by the use of other homogeneous alkaline catalyst

(NaOH or KOH) for transesterification reaction (second stage). Although this method can

utilize unrefined or waste oils for biodiesel production, the process requires multiple reactions,

washing and separation stages. The strong acidic or alkaline catalysts used are highly corrosive

(specially metallurgical), and must be removed from the biodiesel product by multiple washing

steps (Yan et al. 2009). Thus, a significant amount of wastewater is generated, together with a

loss of catalyst.

Figure 1.2 - Global transesterification reaction.

Figure 1.3 - Esterification reaction.

In this regard, recently the heterogeneous catalysts have caught researchers´ attention

because of their advantages over homogenous catalysts in terms of higher biodiesel yield,

higher glycerol purity, easier catalyst separation (simple filtration), lower corrosively and

recoverability, price, safety, more environmentally friendly, and not requiring a washing step

for the crude ester (Agarwal et al. 2012; Wen et al. 2010). Moreover, heterogeneous catalysts

are preferred over homogenous catalysts in biodiesel production from waste cooking oil

(WCO) because saponification reaction is eliminated (Wan Omar et al. 2011). However, it is

worth mentioning that the activity of heterogeneous catalysts is largely reliant on their nature,


6

active sites, structural morphology, porosity, and thermal stability. Heterogeneous catalysts

induce slower reaction rate but the limitation can be resolved by increasing the reaction

temperature and the concentration of reactants (Avhad and Marchetti 2015). Similar to the

homogeneous catalysts, the heterogeneous catalysts are also of two types, namely, acid and

alkaline. Examples of heterogeneous acid catalysts are sulphated metal oxide,

heteropolyacids, acidic ion exchange resin, and sulphonated amorphous catalysts (Sakai et al.

2009). Additional examples of heterogeneous alkaline catalysts include zinc oxide (ZnO),

calcium oxide (CaO), strontium oxide (SrO) (Liu et al. 2007), and titanium oxide (TiO) (Chouhan

and Sarma 2011).

Aiming to overcome the drawbacks/limitations of acid and alkaline catalysts stated before, the

trends of the research in the catalysis technologies for biodiesel production is the development

of solid materials able to simultaneously catalyze the esterification and transesterification

reactions, and the possibility to be re-used. These types of heterogeneous catalysts are called

“bifunctional”, since they act simultaneously as acid and alkaline catalysts. In other words, with

bifunctional catalyst esterification and transesterification are carried out in one single reaction

step and not in two-steps (first esterification and second transesterification) (Farooq et al.

2013; Kondamudi et al. 2011). As stated previously, due to high FFA content in low cost

feedstock, the alkali catalyzed transesterification reaction to produce biodiesel gives low

biodiesel yield because FFA reacts with alkali to form soap, resulting in serious emulsification

and separation problems. To solve this problem, it is possible to produce biodiesel by a two-

steps catalytic process. The process involved the esterification reaction (FFA conversion to

FAME) followed by transesterification (triglycerides conversion to FAME). However, the two-

steps method also faces the problem of catalyst removal in both steps (Enweremadu and

Mbarawa 2009). The heterogeneous bifunctional catalyst would not only eliminate the long

two-steps process but will also reduce the need of high cost equipment; consequently,

minimizing the capital cost of the processing technology for biodiesel production (Alhassan et

al. 2015). Figure 1.4 shows the two types of processes, i.e, the conventional catalysts (two-

step) versus bifunctional catalyst.

THE STUDY PROPOSAL

7

Figure 1.4 - Two-step conventional catalysis versus bifunctional catalyst process.

In line with the circular economy principles, the development of solid catalysts from waste

sources could be a promising way for reducing the environmental burdens of the process and

the production costs. Some research works (e.g., Boey et al. 2009; Farooq et al. 2015) have

focused on the exploitation of waste materials (e.g. shells, ashes, rocks and bones), due to

their abundance and low cost, for solid catalysts preparation. The exploitation of such waste

materials has become very attractive. For example, calcium oxide derived from mollusk shells

(waste mud crab shells) has been proven to be a potential heterogeneous catalyst for biodiesel

production (Viriya-empikul et al. 2010). Chicken and quail eggshells calcined were found to be

a reliable source of CaO, consisting of 85–95 % of calcium carbonate and 15–5 % other

components (MgCO3, phosphate, organic matter and a small amount of metals) in dry

eggshells (Chojnacka 2005). Another potential catalyst with high carbonate originating

naturally and low cost are dolomite (CaMg(CO3)2). The dolomite is a rock that consists of 23.5

% Ca, 12.1 % Mg, 63 % CO3 and 1.4 % other components (Fe, SiO2, PO2, SO4) (Zhao et al. 2019).

Fly ashes are fine particulates that are obtained from the flue gas treatment systems (e.g.

electrostatic precipitators), downstream of the combustion process, in a power plant, using


8

fossil or renewable energetic vector (e.g. coal or wood bark). These ashes are usually rich in

SiO2, Al2O3, CaO, MgO and other oxides, which are pointed out to have excellent catalytic

properties for transesterification and esterification reactions (Ho et al. 2014; Sharma et al.

2012).

Globally, the cost of production has been the main barrier in commercializing biodiesel. In the

literature, it is consensual that the oily feedstock is the major contributor, about 80 % (Mansir

et al. 2018), for the total production costs. The use of edible oils sparks concern in terms of

food security while the non-edible oils need additional pretreatment steps. On the other hand,

the wide availability of edible oils guarantees the supply while the alternative of non-edible

oils is subject to an intermittent supply due to difficulties in its collection (Nurfitri et al. 2013).

The waste cooking oils (WCO) are some vegetable oils that have been previously used for frying

or cooking and can constitute an additional source of raw material for biodiesel production.

This feedstock can be two to three times cheaper than virgin vegetable oils (Demirbas 2009)

and it is important to emphasize that the use of WCO can reduce biodiesel production costs

by 60–90 % (Zhang et al. 2003). The WCO has many advantages, such as: (i) abundant supply,

(ii) relatively inexpensive, and (iii) environmental benefits (e.g., recovery of hard management

waste).

In many countries the waste vegetable oils are abundant; they can be readily collected from

households and HORECA sector (hotels restaurants and catering). Nurfitri et al. (2013)

reported that, annually, U.S recorded 0.7–1.0 million tons of collected waste oil, Turkey

350.000 tons, Canada 120.000 tons and Bogotá about 4000 tons. In addition, there are

unknown amounts of uncollected oils which are discarded through sinks, regular garbage and

eventually seeps into the soil and water sources. Furthermore, it is generally accepted that

reusing used cooking oil for human consumption is harmful to health (Wei See et al. 2006).

The main factors affecting the yield in FAME production, must be chosen judiciously, namely

(Leung et al. 2010; Ni and Meunier 2007): alcohol type, molar ratio of alcohol to oil, catalyst

loading, reaction temperature, reaction time and type of reactor. Therefore, to find the

THE STUDY PROPOSAL

9

relationship between the operating conditions (levels of the factors evaluated) and the best

yield to FAME requires specialized optimization methods. Response Surface Methodology

(RSM) based on a Box–Behnken experimental design is an option, that corresponds to a set of

mathematical and statistical techniques employed for designing experiments, creating

correlations (regression model), evaluating the effects of several factors, and their interaction

effects for desirable responses (Liu et al. 2014; Salamatinia et al. 2010). There is commercial

software (e.g. Design – Expert) to assist this statistical data processing and analysis.

It is known that the catalytic behavior depends on the morphological characteristics of the

solid material, because the catalytic process takes place on its surface (outer and inner). The

most utilized techniques to characterize materials’ morphology are BET, SEM and EDX. In terms

of physical properties, the surface area is the place of catalytic activity, but only a part is utilized

in the catalytic reaction (active center or site). In basic and acid catalysts, the active sites not

only occupy a little fraction of the surface, but also differ in basic and acid strength and

sometimes in nature. Hammett indicators are often used to determine the acid and basic

strengths of a material. FTIR is useful to identify the main chemical functional groups present

on the surface of solid materials. For bulk properties X-ray diffraction (XRD) is used to find: (i)

the crystalline phases, (ii) crystalline degree and (iii) crystallite size and superficial elemental

atomic concentration by means of X-ray photoelectronic spectroscopy (XPS). This technique is

frequently used to assist the interpretation of the catalytic activity of one material.

Colombia has the fourth largest production of palm oil worldwide with 1.309.586 tons per year

with 550.200 hectares of palm plantation, ranking the third in production of biodiesel (518.745

tons per year) in South America, followed by Argentina and Brazil. It is noteworthy that all

biodiesel production in Colombia is from palm oil (edible feedstock). This sparks concern in

terms of food security in addition to the misuse of fertile lands, water footprint and

deforestation. So far, the costs of refined palm oil (RPO) feedstock and chemicals (methanol

+ NaOH) involved in biodiesel production reaches 88 % of the selling price of this biofuel in

Colombia; therefore the government must subsidize this industry (80 million dollars per year)

(Acevedo et al. 2015). That is why using the exploration of feedstock and catalysts from waste


10

should be intensified in order to make the biodiesel production process low cost, affordable

and sustainable.

During the last years, the Colombian government has promoted the production and use of

biofuels as a part of a strategy to reduce the petroleum dependence and to support the rural

development. Colombian fuels must contain a percent of bioethanol and biodiesel blended

with gasoline and diesel, respectively. Particularly, for biodiesel and according to the policy

act 2629 of 2007, B10 blends (10 % biodiesel and 90 % diesel) must be implemented since 2010

and B20 since 2012. Nevertheless, these blending targets were not reached in these years.

However, the market growth in Colombia implicates the expansion on biodiesel demand from

B10 to B20 blend for 2020 (Rincón et al. 2015). In addition, the efforts to employ waste need

to receive the support of governments, for instance, in terms of tax relief/reduction. Most

importantly, the government can play a serious role in introducing regulations to enforce the

use of waste substances. In this way, it is no longer a choice but a must to include the intake

materials from waste products in many more processes.

1.2 Thesis objectives and structure

This research work aims to prepare efficient bifunctional solid catalysts, from waste materials,

to produce FAME from mixtures of low cost vegetable oils (WCO and RPO) with methanol, for

later implementation in both optimized batch and continuous production processes. Thus,

through the use of waste materials it could be possible a reduction the production costs of

biodiesel and simultaneously extend the life-cycle of the materials in the economy, i.e.,

promoting a circular economy.

Besides that main objective, this proposal also comprises the following specific objectives:

- Prepare and characterize feedstocks (mixtures of oils – WCO and RPO), in terms of

some physical chemical properties.

THE STUDY PROPOSAL

11

- Prepare and characterize solid catalysts from waste, using preparation procedures and

instrumental analytical techniques, respectively.

- Assess the activity and stability of the solid catalysts prepared, using the yield to FAME

and the conversion of the FFA as dependent target variables.

- Optimize the production of FAME using the Response Surface Methodology (RSM)

based on a Box–Behnken experimental design for most efficient bifunctional catalyst

(using the main factors affecting yield to FAME) and evaluating in both batch reactor

and in continuous fixed bed reactor.

The novelty of this work is the successful use of a residual materials (either catalyst or

feedstock) in the FAME production, carrying out the transesterification and esterification

reactions simultaneously in only one stage, innovating over the conventional production

process that requires two stages.

For better organization and understanding, the thesis was structured in five sections, as shown

in Figure 1.5. Section A presents the study, introducing to the topic, which includes the

motivation, the relevance of the study, the objectives and the structure of the thesis.

In sections B to D (in scientific article format), the studies carried out to achieve the objectives

of the work are presented. It started with the selection of a catalyst with a bifunctional

character, prepared with residual materials. The catalytic performance in the FAME production

was assessed and one catalyst (out of eight) was selected. Then, an optimization study was

carried out in batch process for FAME production, using the catalyst selected in the previous

study. A third study was conducted, this time to produce FAME in a continuous process with

the selected catalyst. Both optimization studies made it possible to identify the most significant

(in the tested range) operating variables. Each of these experimental works provided results

that allowed progressively to advance in the achievement of the objectives defined for this

thesis.

Therefore, Section B presents the study to evaluate and choose an efficient bifunctional solid

catalyst from waste materials (biomass fly ash, dolomite, egg shells and polyethylene


12

terephthalate (PET) from waste plastic containers) for production of FAME with methanol,

using mixtures of refined palm oil (RPO) and used cooking oil (WCO) in different mass ratios.

This study was carried out in a batch reactor due to the ease of operation and control of the

process variables. More, the batch reactor allows to carry out the tests of catalytic activity in

a reliable and efficient way.

Section C presents the study of optimization of the FAME production process with methanol,

using mixtures of WCO and RPO and the catalyst (powder) selected in the previous work

(section B). The Response Surface Methodology (RSM) based on a Box–Behnken experimental

design was used to test four operating variables (in a batch reactor), namely: catalyst loading,

methanol/oil molar ratio, RPO/WCO mass ratio and reaction temperature. Additionally, it was

carried out a study of reusability of the catalyst aiming to assess its performance and catalytic

stability over several cycles of utilization in the FAME synthesis process. The optimal operating

conditions found by the regression model were used in this assay.

Section D presents the design, construction of a continuous fixed bed reactor (novel and simple

reaction system to produce FAME with solid catalyst) and optimization study. The RSM based

on a Box–Behnken experimental design was used to test the influence of three operating

variables, namely: methanol/oil molar ratio, RPO/WCO mass ratio and residence time. It is

important to mention that the biomass fly ash catalyst is a very fine powder and to be

evaluated in a continuous fixed-bed reactor it was necessary to pelletize it. Additionally, the

catalytic stability of the pelletized biomass fly ash was evaluated through an assay during 32 h

of operation in the continuous fixed bed reactor. The optimal operating conditions found by

the regression model were used in this assay.

Finally, in Section E, final remarks are made about results obtained throughout the study, and

proposals for future studies are presented.

THE STUDY PROPOSAL

13

1.3 Scientific production within the scope of the thesis

The global information of the studies presented in Sections B to D of this thesis were published

or submitted in scientific journals with peer review or at conferences, as mentioned in the next

subchapters.

1.3.1 Articles published in scientific journals

E. M. Vargas, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Solid catalysts obtained

from wastes for FAME production using mixtures of refined palm oil and waste cooking

oils,” Renew. Energy, vol. 136, pp. 873–883, 2019, doi: 10.1016/j.renene.2019.01.048.

E. M. Vargas, J. L. Ospina, L. A. C. Tarelho, and M. I. Nunes, “FAME production from

residual materials: Optimization of the process by Box–Behnken model,” Energy

Reports, vol. 6, pp. 347–352, 2019, doi: 10.1016/j.egyr.2019.08.071.

E. M. Vargas, J. L. Ospina, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Optimization

of FAME production from blends of waste cooking oil and refined palm oil using

biomass fly ash as a catalyst”, Renew. Energy, vol. 163, pp. 1637–1647, 2021, doi:

https://doi.org/10.1016/j.renene.2020.10.030.

E. M. Vargas, Duvan O. Villamizar, M. C. Neves, and M. I. Nunes, “Pelletized biomass fly

ash for FAME production: optimization of a continuous process”, Fuel, vol. 293, pp.

120425, 2021, doi: https://doi.org/10.1016/j.fuel.2021.120425.

1.3.2 Articles published in scientific conferences

E. M. Vargas, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Produção de biodiesel

usando materiais residuais”, conferência nacional: Green Business Week, 15 a 17 de

março de 2017 Lisboa, Portugal.

E. M. Vargas, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Biodiesel production using

residual materials”, conferência internacional: 4th international conference, Wastes

Solutions Treatments Opportunities, 25 e 26 de setembro de 2017 Porto, Portugal.


14

E. M. Vargas, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Produção de biodiesel a

partir de materiais residuais”, TechDays, workshops e conferência, 12 a 14 de outubro

de 2017 Aveiro, Portugal.


residual materials: Optimization of the process by Box–Behnken model,” 6th

International Conference on Energy and Environment Research, ICEER 2019, 22-25 July,

University of Aveiro, Portugal.

1.3.3 Participation in scientific events

Perspectivas del post-conflicto: una mirada desde la ingeniería, Bogotá. Título do

trabalho: Biodiesel production using residual materials: an alternative for Colombia. 9

e 10 outubre de 2017, Bogotá Colombia.

4to encuentro de semilleros de investigación, ciencia, arte e innovación. Título do

trabalho: Producción de biodiésel a partir de mezclas de aceite de palma (RBD) con

aceites usados de cocina (AUC), utilizando como catalizador CaO-SiO2. 2 e 3 de

novembro de 2017, Bogotá Colombia.

UA Open Campus, Aveiro. Título do trabalho: Produção de biodiesel a partir de

materiais residuais. 22 a 24 de março de 2018.

4to encuentro de semilleros de investigación, ciencia, arte e innovación. Título do

trabalho: Producción de biodiésel a partir de mezclas de aceite de palma (RBD) con

aceites usados de cocina (AUC), utilizando como catalizador CaO-SiO2. 2 e 3 de

novembro de 2017, Bogotá Colombia.

Perspectivas del post-conflicto: una mirada desde la ingeniería. Título do trabalho:

Biodiesel production using residual materials: an alternative for Colombia. 9 e 10

outubre de 2017, Bogotá Colombia.

THE STUDY PROPOSAL

15

Figure 1.5 - Structure of the thesis in five sections.

Section A

The study proposal and presentation of the thesis

• Introduction (considers the conceptual framework,justification and relevance of the work)

• Objectives

• Thesis structure

Section B

Solid bifunctional catalyst selection

• Produce, evaluate and choose an efficient solid bifunctional catalyst fromresidual materials (biomass fly ashes, dolomite, eggshells and polyethyleneterephthalate (PET) of waste plastic packaging) for FAME production usingmethanol, and mixtures of refined palm oil (RPO) and waste cooking oil(WCO) in different mass ratios.

Section C

Optimization of FAME production in batch mode

operation

• Presents the study to optimize the FAME production process with methanol,using mixtures of WCO and RPO and biomass fly ashes as chosen catalyst;using the Response Surface Methodology (RSM) based on a Box–Behnkenexperimental design (four operating variables, that are: catalyst loading,methanol/oil molar ratio, RPO/WCO mass ratio and reaction temperature) ina batch reactor. Additionally, it was carried out a study of reusability of thecatalyst.

Section D

Optimization of FAME production in a

continuous fixed-bed reactor

• Presents the design and construction of a continuous fixed bed reactor(CFBR) and study to optimize the FAME production process with methanol,using mixtures of WCO and RPO and biomass fly ashes as chosen catalyst;using the Response Surface Methodology (RSM) based on a Box–Behnkenexperimental design (three operating variables, that are: residence time,WCO/RPO mass ratio and methanol/oil molar ratio) in a continuous fixedbed reactor. Additionally, the catalytic stability of the catalyst evaluatedthrough an assay during 32 h of operation.

Section E

Final remarks

• Main conclusions of the study and suggestions for future work in theproduction of biodiesel using waste as raw materials.


16

References

Acevedo, Juan C, Jorge a Hernández, Carlos F Valdés, and Samir Kumar Khanal. 2015. “Analysis of Operating Costs for Producing Biodiesel from Palm Oil at Pilot-Scale in Colombia.” Bioresource technology 188: 117–23. http://www.ncbi.nlm.nih.gov/pubmed/25660089 (April 8, 2015).

Agarwal, Madhu, Garima Chauhan, S. P. Chaurasia, and Kailash Singh. 2012. “Study of Catalytic Behavior of KOH as Homogeneous and Heterogeneous Catalyst for Biodiesel Production.” Journal of the Taiwan Institute of Chemical Engineers 43(1): 89–94. http://dx.doi.org/10.1016/j.jtice.2011.06.003.

Alhassan, Fatah H., Umer Rashid, and Y.H. Taufiq-Yap. 2015. “Synthesis of Waste Cooking Oil-Based Biodiesel via Effectual Recyclable Bi-Functional Fe2O3MnOSO42−/ZrO2 Nanoparticle Solid Catalyst.” Fuel 142: 38–45. http://linkinghub.elsevier.com/retrieve/pii/S0016236114010333 (April 9, 2015).

Avhad, M. R., and J. M. Marchetti. 2015. “A Review on Recent Advancement in Catalytic Materials for Biodiesel Production.” Renewable and Sustainable Energy Reviews 50: 696–718. http://dx.doi.org/10.1016/j.rser.2015.05.038.

Boey, Peng Lim, Gaanty Pragas Maniam, and Shafida Abd Hamid. 2009. “Biodiesel Production via Transesterification of Palm Olein Using Waste Mud Crab (Scylla Serrata) Shell as a Heterogeneous Catalyst.” Bioresource Technology 100(24): 6362–68. http://dx.doi.org/10.1016/j.biortech.2009.07.036.

Chojnacka, Katarzyna. 2005. “Biosorption of Cr(III) Ions by Eggshells.” Journal of Hazardous Materials 121(1–3): 167–73.

Chouhan, a.P. Singh, and a.K. Sarma. 2011. “Modern Heterogeneous Catalysts for Biodiesel Production: A Comprehensive Review.” Renewable and Sustainable Energy Reviews 15(9): 4378–99. http://linkinghub.elsevier.com/retrieve/pii/S1364032111003595 (July 16, 2014).

Demirbas, Ayhan. 2009. “Biodiesel from Waste Cooking Oil via Base-Catalytic and Supercritical Methanol Transesterification.” Energy Conversion and Management 50(4): 923–27. http://dx.doi.org/10.1016/j.enconman.2008.12.023.

Enweremadu, C. C., and M. M. Mbarawa. 2009. “Technical Aspects of Production and Analysis of Biodiesel from Used Cooking Oil-A Review.” Renewable and Sustainable Energy Reviews 13(9): 2205–24.

Farooq, Muhammad, and Anita Ramli. 2015. “Biodiesel Production from Low FFA Waste Cooking Oil Using Heterogeneous Catalyst Derived from Chicken Bones.” Renewable Energy 76: 362–68. http://dx.doi.org/10.1016/j.renene.2014.11.042.

Farooq, Muhammad, Anita Ramli, and Duvvuri Subbarao. 2013. “Biodiesel Production from Waste Cooking Oil Using Bifunctional Heterogeneous Solid Catalysts.” Journal of Cleaner Production 59: 131–40. http://linkinghub.elsevier.com/retrieve/pii/S095965261300396X (May 14, 2015).

THE STUDY PROPOSAL

17

Ho, Wilson Wei Sheng, Hoon Kiat Ng, Suyin Gan, and Sang Huey Tan. 2014. “Evaluation of Palm Oil Mill Fly Ash Supported Calcium Oxide as a Heterogeneous Base Catalyst in Biodiesel Synthesis from Crude Palm Oil.” Energy Conversion and Management 88: 1167–78. http://linkinghub.elsevier.com/retrieve/pii/S0196890414002623 (March 7, 2015).

Kiss, Anton A., Alexandre C. Dimian, and Gadi Rothenberg. 2008. “Biodiesel by Catalytic Reactive Distillation Powered by Metal Oxides.” Energy and Fuels 22(1): 598–604.

Kondamudi, Narasimharao, Susanta K. Mohapatra, and Mano Misra. 2011. “Quintinite as a Bifunctional Heterogeneous Catalyst for Biodiesel Synthesis.” Applied Catalysis A: General 393(1–2): 36–43. http://linkinghub.elsevier.com/retrieve/pii/S0926860X10007799 (July 1, 2015).

Lam, Man Kee, Keat Teong Lee, and Abdul Rahman Mohamed. 2009. “Sulfated Tin Oxide as Solid Superacid Catalyst for Transesterification of Waste Cooking Oil: An Optimization Study.” Applied Catalysis B: Environmental 93(1–2): 134–39.

Leung, Dennis Y C, Xuan Wu, and M. K H Leung. 2010. “A Review on Biodiesel Production Using Catalyzed Transesterification.” Applied Energy 87(4): 1083–95. http://linkinghub.elsevier.com/retrieve/pii/S0306261909004346 (July 10, 2014).

Liu, Wei, Ping Yin, Xiguang Liu, and Rongjun Qu. 2014. “Design of an Effective Bifunctional Catalyst Organotriphosphonic Acid-Functionalized Ferric Alginate (ATMP-FA) and Optimization by Box-Behnken Model for Biodiesel Esterification Synthesis of Oleic Acid over ATMP-FA.” Bioresource Technology 173: 266–71. http://dx.doi.org/10.1016/j.biortech.2014.09.087.

Liu, Xuejun, Huayang He, Yujun Wang, and Shenlin Zhu. 2007. “Transesterification of Soybean Oil to Biodiesel Using SrO as a Solid Base Catalyst.” Catalysis Communications 8(7): 1107–11.

Mansir, Nasar et al. 2018. “Modified Waste Egg Shell Derived Bifunctional Catalyst for Biodiesel Production from High FFA Waste Cooking Oil. A Review.” Renewable and Sustainable Energy Reviews 82(November 2016): 3645–55.

Ni, J., and F. C. Meunier. 2007. “Esterification of Free Fatty Acids in Sunflower Oil over Solid Acid Catalysts Using Batch and Fixed Bed-Reactors.” Applied Catalysis A: General 333(1): 122–30.

Nurfitri, Irma et al. 2013. “Potential of Feedstock and Catalysts from Waste in Biodiesel Preparation: A Review.” Energy Conversion and Management 74: 395–402. http://linkinghub.elsevier.com/retrieve/pii/S0196890413002586 (March 13, 2015).

Rincón, Luis E. et al. 2015. “Optimization of the Colombian Biodiesel Supply Chain from Oil Palm Crop Based on Techno-Economical and Environmental Criteria.” Energy Economics 47: 154–67. http://linkinghub.elsevier.com/retrieve/pii/S0140988314002631 (February 3, 2015).


18

Sakai, Tsutomu, Ayato Kawashima, and Tetsuya Koshikawa. 2009. “Economic Assessment of Batch Biodiesel Production Processes Using Homogeneous and Heterogeneous Alkali Catalysts.” Bioresource Technology 100(13): 3268–76. http://dx.doi.org/10.1016/j.biortech.2009.02.010.

Salamatinia, Babak, Hamed Mootabadi, Subhash Bhatia, and Ahmad Zuhairi Abdullah. 2010. “Optimization of Ultrasonic-Assisted Heterogeneous Biodiesel Production from Palm Oil: A Response Surface Methodology Approach.” Fuel Processing Technology 91(5): 441–48.

Sharma, Meeta, Arif Ali Khan, S. K. Puri, and D. K. Tuli. 2012. “Wood Ash as a Potential Heterogeneous Catalyst for Biodiesel Synthesis.” Biomass and Bioenergy 41: 94–106. http://dx.doi.org/10.1016/j.biombioe.2012.02.017.

Viriya-empikul, N. et al. 2010. “Waste Shells of Mollusk and Egg as Biodiesel Production Catalysts.” Bioresource Technology 101(10): 3765–67. http://dx.doi.org/10.1016/j.biortech.2009.12.079.

Wan Omar, Wan Nor Nadyaini, and Nor Aishah Saidina Amin. 2011. “Optimization of Heterogeneous Biodiesel Production from Waste Cooking Palm Oil via Response Surface Methodology.” Biomass and Bioenergy 35(3): 1329–38. http://linkinghub.elsevier.com/retrieve/pii/S0961953410005076 (May 29, 2015).

Wei See, Siao, Sathrugnan Karthikeyan, and Rajasekhar Balasubramanian. 2006. “Health Risk Assessment of Occupational Exposure to Particulate-Phase Polycyclic Aromatic Hydrocarbons Associated with Chinese, Malay and Indian Cooking.” Journal of Environmental Monitoring 8(3): 369–76.

Wen, Zhenzhong et al. 2010. “Biodiesel Production from Waste Cooking Oil Catalyzed by TiO2-MgO Mixed Oxides.” Bioresource Technology 101(24): 9570–76. http://dx.doi.org/10.1016/j.biortech.2010.07.066.

Yan, Shuli, Steven O. Salley, and K. Y. Simon Ng. 2009. “Simultaneous Transesterification and Esterification of Unrefined or Waste Oils over ZnO-La2O3 Catalysts.” Applied Catalysis A: General 353(2): 203–12.

Zhang, Y, M.a Dubé, D.D McLean, and M Kates. 2003. “Biodiesel Production from Waste Cooking Oil: 2. Economic Assessment and Sensitivity Analysis.” Bioresource Technology 90(3): 229–40. http://linkinghub.elsevier.com/retrieve/pii/S0960852403001500 (December 26, 2014).

Zhao, Shuang et al. 2019. “Experimental Investigation on Biodiesel Production through Transesterification Promoted by the La-Dolomite Catalyst.” Fuel 257(July): 116092. https://doi.org/10.1016/j.fuel.2019.116092.

19

SECTION B – Bifunctional solid catalyst selection

This section aims to develop an efficient bifunctional solid catalyst that would be able to

catalyze the esterification and transesterification reactions simultaneously (single reaction

step). These reactions were performed in batch regime using methanol, mixtures of waste

cooking oil and refined palm oil and solid catalysts. The solid catalysts were prepared from

waste feedstocks (biomass fly ashes, natural dolomite rock, chicken eggshells and polyethylene

terephthalate - PET of waste plastic packaging) and characterized by SEM, EDX, XRD, BET, FT-

IR and Hammett indicators. Eight potential solid catalysts were prepared and tested. Their

performance in the esterification and transesterification catalysis was assessed by the FFA

conversion and the yield to FAME (for both reactions), respectively.

The information presented in this section was adapted from the following published article:

E. M. Vargas, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Solid catalysts obtained

from wastes for FAME production using mixtures of refined palm oil and waste cooking

oils,” Renew. Energy, vol. 136, pp. 873–883, 2019, doi: 10.1016/j.renene.2019.01.048.


20

BIFUNCTIONAL SOLID CATALYST SELECTION

21

2 Solid catalysts obtained from wastes for FAME production using

mixtures of refined palm oil and waste cooking oils

Abstract: More than 95 % of biodiesel production feedstocks come from edible oils, however

it may cause some problems such as the competition for land use for food production

and biodiesel production. The waste cooking oils (WCO) are an alternative feedstock for

biodiesel production; its usage reduces significantly the cost of biodiesel production and has

environmental benefits, e.g., a waste recovery instead of its elimination. This work aims to

produce a low-cost efficient solid catalyst for fatty acid methyl esters (FAME) production using

mixtures of RPO and WCO (M1 (100 % RPO), M2 (75 % RPO, 25 % WCO), M3 (50 % RPO, 50 %

WCO), M4 (25 % RPO, 75 % WCO) and M5 (100 % WCO)). Four low cost catalysts were prepared

(biomass fly ashes, natural dolomite rock, chicken eggshells and polyethylene terephthalate -

PET) and their derivatives, characterized (by SEM, EDX, XRD, BET, FT-IR and Hammett

indicators) and tested regarding their performance in FAME production. The maximum yield

of FAME achieved was around 96 wt% for biomass fly ashes catalyst at 60 oC, 9:1 (mol/mol) of

methanol to oil mixture,10 wt% catalyst to oil mixture, 600 rpm, over 180 min in batch reactor.

The results point out for promising bifunctional catalysts able to yield also conversion of free

fatty acids up to 100 % using mixtures of RPO and WCO.

Keywords: Waste materials; FAME; solid catalyst; bifunctional catalysts; waste cooking oil;

refined palm oil.


22

2.1 Introduction

Energy is a basic requirement for human existence and its demand grows every year. In the

current situation, the foremost amount of primary and useful energy is supplied by the

conventional fossil fuel resources, such as gasoline, liquefied petroleum gas, diesel fuel, and

natural gas. However, the use of fossil fuels has several influences on the environment, such

as large greenhouse gas emissions, acid rain, resources depletion, etc. In addition to serious

environmental issues, dwindling reserves of crude oil, fluctuating petroleum fuel prices, have

made today's need to find alternative “green” sources of energy, which are sustainable,

environmentally compatible, economically competitive, and easily available. One of the most

promising sources is biodiesel, an alternative diesel fuel derived from renewable sources with

high quality, which allows the replacement of fossil diesel oil (Leung et al. 2010). Usually the

biodiesel production is a catalyzed process where alkali or acid compounds are used,

respectively, for the conversion of triglycerides (transesterification reaction) and free fatty

acids – FFA (esterification reaction) into fatty acid methyl esters (FAME), when methanol is

used in the synthesis. In line with the circular economy principles, the development of solid

catalysts from waste sources could be a promising way for reducing the environmental

burdens of the process and the production costs. Some research works (Boey et al. 2009;

Farooq et al. 2015) have focused on the exploitation of waste materials (e.g. shells, ashes, rocks

and bones), due to their abundance and low cost, for solid catalysts preparation.



et al. 2018), for the total production costs. The use of edible oils sparks concern in terms of

food security while the non-edible oils need additional pre-treatment steps. On the other

hand, the wide availability of edible oils guarantees the supply while the alternative of non-

edible oils is subject to an intermittent supply (Nurfitri et al. 2013).


23

The waste cooking oils (WCO) are some vegetable oils that have been previously used for frying

or cooking and can constitute an additional source of raw material for biodiesel production.

This feedstock can be two to three times cheaper than virgin vegetable oils (Demirbas 2009).

Furthermore, it is generally accepted that reusing used cooking oil for human consumption is

harmful to health (Chen et al. 2012) and the WCO is difficult to manage.

In the research that focuses on the biodiesel production process there is a new trend of

bifunctional heterogeneous catalysts that, if used properly, can catalyze both

transesterification and esterification reactions simultaneously. This ability is due to the

presence of both basic and acidic sites on the same catalysts. Additionally, this kind of catalysts

can be modified to introduce/improve certain physicochemical properties needed to handle

with low grade feedstocks (e.g. some WCO). Usually these raw-materials have high FFA and/or

water contents, which are undesirable for transesterification reaction (Farooq et al. 2013).

This work aims to produce an efficient solid bifunctional catalyst from residual materials for

FAME production using mixtures of refined palm oil (RPO) and WCO in different ratios. Thus,

this work tackles two current environmental concerns giving an alternative for recovering

some important wastes fluxes (WCO, fly ashes, dolomite, eggshells and polyethylene

terephthalate (PET) plastic garbage containers), aligned with the principles of circular

economy, and for reducing the dependence of fossil fuel, through the production of biofuel

mainly from waste materials feedstocks.

2.2 Materials and Methods

Solid catalysts were prepared and characterized in terms of some of their chemical, physical

and structural properties. The raw-material for FAME synthesis consisted of a mixtures of WCO

and RPO in different ratios. The adopted procedures are described in next sections. The

experimental plan for assessing the performance of the catalysts in the FAME production will

be presented later as well as the analytical methods used.


24

2.2.1 Materials

Waste cooking oil for FAME production was provided by a local collecting company (Bioils) in

Bogotá, Colombia. The WCO was pre-treated by filtration and heating (at 110 °C for 1 h) to

remove suspended particles and traces of water, respectively. The RPO was purchased at a

local store in Bogotá. The solid waste materials for catalyst preparation were obtained from

the following sources:

Biomass fly ashes – collected at the electrostatic precipitator of a thermal power-plant

using residual forest biomass (derived from eucalyptus) as fuel, located in the Centre

Region of Portugal;

Natural dolomite rock- the mining industry in Colombia;

Eggshells- several restaurants of Bogotá;

PET- plastic bottles picked from garbage containers at the University Jorge Tadeo

Lozano, Bogotá.

All the chemicals used were of analytical grade except n-hexane (GC grade) and methyl

heptadecanoate (analytical standard) from Sigma-Aldrich and Merck.

2.2.2 Oil mixtures (RPO and WCO) characterization

Five feedstocks were prepared using different mass ratios of RPO and WCO: M1 (100 % RPO),

M2 (75 % RPO, 25 % WCO), M3 (50 % RPO, 50 % WCO), M4 (25 % RPO, 75 % WCO) and M5

(100 % WCO).

These mixtures were characterized in terms of: acid value (NTC 218 (ICONTEC 218 1999)),

density (NTC 336 (ICONTEC 2011)), saponification number (NTC 335(ICONTEC 2011)), viscosity

(ASTM D445) (ASTM 2010), and moisture content (Karl Fisher, Coulometer 831-Metrohm). The

saponification number (𝑆𝑁) was used to calculate the molecular mass (𝑀𝑊) according to

Equation 2.1 (Mansir et al. 2018).

𝑀𝑊 =56,1×1000×3

𝑆𝑁 (2.1)


25

The FFA content was calculated from the acid value (𝐴𝑉, mg KOH/g) using Equation 2.2 (Mansir

et al. 2018).

FFA =𝐴𝑉

2 (2.2)

2.2.3 Catalysts preparation and characterization

Eight catalysts were prepared using low cost feedstocks, implementing the procedures

summarized in Table 2.1. The sulfonation or the addition of silicon to some catalysts aimed to

enhance their acid strength.

Table 2.1 - Solid catalysts preparation procedures.

Catalyst reference Preparation procedure

Biomass fly ash

FAD Dry at 120 °C for 5 h sieve 75 μm.

FAC Calcine FAD at 700 °C for 5 h sieve 75 μm.

Natural dolomite rock

Dolomite C Mill and sieve at 45 µm and calcine at 800 °C for 2 h.

Dolomite CSC Impregnate dolomite C with H2SO4 2M for 6 h at room temperature. Then, filter and dry for 12 h at 110 °C. Finally, calcine at 500 °C for 4 h.

Eggshells

CaO-SiO2

Wash with water and dry at 120 °C for 3 h. Then, sieved at 63 μm and calcine at 800 °C for 4 h. Impregnate with Na2SiO3 0.4 M aqueous solution at room temperature for 4 h. Finally dry at 100 °C for 12 h and calcine at 800 °C for 4 h.

CaO-S-SiO2 Impregnate CaO-SiO2 with H2SO4 2 M for 6 h at room temperature, dry at 110 °C for 12 h and calcine at 500 °C for 3 h.

PET

CA-PET

Reduce (cut) the PET containers to small pieces (<1 cm2) and heat 10 °C/min for 2 h from room temperature up to 450 oC under a nitrogen atmosphere. Impregnate the resulting product with H2SO4 98 % (1.5:1 v/w, H2SO4: PET) at 150 °C for 2 h. Then, wash with water and dry at 120°C for 6 h. Mill and sieve at 106 μm, and finally dry at 105 °C for 24 h.

CA-PET-S

Impregnate CA-PET with fuming sulfuric acid (5:1 v/w, H2SO4:CA-PET) at 150 °C for 10 h under a nitrogen atmosphere. Wash with water (until no sulfate ions are detected, using turbidimetric method) and dry at 105 °C for 24 h.


26

The solid catalysts were characterized in terms of: (i) crystallographic structures, by powder X-

ray (XRD, PAN analytical Empyrean X-ray diffractometer equipped with Cu-Kα radiation source

λ = 1.54178 Å at 45 kV/ 40 mA); (ii) surface area, pore size and pore volume, by Brunauer-

Emmet-Teller sorption isotherm (BET, using N2 at -196 oC in Micromeritics ASAP 2020);

(iii) surface morphology and quantitative elemental composition analysis, by surface scanning

electron microscopy (SEM, using FEG-SEM Hitachi S4100 microscope operated at 25 kV) and

energy dispersive X-ray spectroscopy (EDX, using a HR-FESEM Hitashi SU-70 operated at 15 kV,

equipped with a Bruker Quantax 400 EDS system); (iv) surface functional species by Fourier

transform infrared (FTIR, Agilent CARY 630 with wave number range from 400 to 4000 cm-1);

and (v) basic and acid strength by using Hammett indicators (indicators for basic strength:

neutral red (pKa = 6.8), bromothymol blue (pKa = 7.2), phenolphthalein (pKa = 9.3), indigo

carmine (pKa = 12.2) and 2,4-dinitroaniline (pKa = 15.0); indicators for acid strength:

bromothymol blue (pKa = 7.2), neutral red (pKa = 6.8), bromocresol purple (pKa = 6.1),

bromocresol green (pKa = 4.7) and bromophenol blue (pKa = 3.8)). The latter method was

carried out by dispersing about 25 mg of the sample in 5.0 mL of a solution of Hammett

indicators (0.5 mg of indicator in 10 mL of methanol for basic strength or 10 mL of benzene for

acid strength), and left for 2 h in order to attain the equilibrium. After reaching equilibrium,

the color on the catalyst and solution were identified.

2.2.4 FAME synthesis

The experiments for FAME production were carried out in batch reactor (in stainless steel, 1 L

of capacity, equipped with temperature control and mechanical agitator) at 60 oC, 9:1

(mol/mol) of methanol to oil mixture, 10 wt% catalyst to oil mixture, 600 rpm and over 180

min. After the pre-defined reaction time, for each assay, the catalyst and methanol were

separated from the reaction mixture by centrifugation and evaporation, respectively. Then,

the supernatant was placed into a separating funnel over 12 h for phase separation. The upper

layer was dried with anhydrous sodium sulfate (10 wt%) and weighed. The resulting mixture,

hereafter is so-called purified final mixture, was analyzed by gas chromatography for FAME


27

determination and was titrated with a KOH solution for final acid value quantification

(ICONTEC 218 1999).

The Shimadzu G-C 2014 chromatograph was equipped with a flame ionization detector and a

capillary column SGEBP-20 60 m x 0.25 mm i.d. x 0.25 µm film thickness with a stationary phase

of polyethylene glycol; the carrier gas was helium with a flow rate of 16.7 mL/min and a

pressure of 2.5 atm; the injector (AOC-20i) was operated at 200 o C and an injection volume of

2.0 µL in Split mode. Methyl heptadecanoate was used as internal standard and hexane as the

solvent. The content of methyl esters was calculated based on the standard method UNE-EN

ISO 14103:2011 (AENOR-EN 14103 2011) and expressed as concentration of FAME using the

Equation 2.3.

𝐶 =∑ 𝐴−𝐴𝐸𝐼

𝐴𝐸𝐼×

𝑊𝐸𝐼

𝑊 (2.3)

Where 𝐶 is the concentration of FAME in the purified final mixture (w/w), ∑ 𝐴 is the total peak

areas of the methyl ester from C14 until C24:1, 𝐴𝐸𝐼 is the peak area corresponding to methyl

heptadecanoate, 𝑊𝐸𝐼 is the mass (mg) of methyl heptadecanoate used and 𝑊 is the mass (mg)

of the sample used in the analysis.

The catalysts performance was expressed in terms of FAME yield, Equation 2.4, and FFA

conversion, Equation 2.5 (Uprety et al. 2016; Wan Omar et al. 2011a).

FAME yield (%) =𝐶 ×Total mass of purified final mixture

Mass of oil used in the experiment × 100 (2.4)

FFA conversion (%) = (1 −𝐴𝑉𝑓

𝐴𝑉𝑖) × 100 (2.5)

Where 𝐴𝑉𝑖 and 𝐴𝑉𝑓 correspond to the acid value of the initial oil mixture and of the purified

final mixture, respectively.


28

2.3 Results and discussion

2.3.1 Oil mixtures characterization

The results of the characterization of the oil mixtures prepared for this study are shown in

Table 2.2.

Table 2.2 - Properties of the oil mixtures used.

Mixture reference

M1 M2 M3 M4 M5

%WCO 0 25 50 75 100

%RPO 100 75 50 25 0

Moisture (wt%)

0.067 ± 0.010 0.141 ± 0.017 0.170 ± 0.003 0.177 ± 0.013 0.197 ± 0.012

Density (g/mL)

0.908 ± 0.008 0.907 ± 0.004 0.913 ± 0.010 0.905 ± 0.007 0.906 ± 0.003

AV (mg KOH/g)

0.307 ± 0.004 1.249 ± 0.061 2.458 ± 0.082 3.873 ± 0.088 4.934 ± 0.252

FFA (wt%)

0.172 ± 0.0048 0.622 ± 0.0512 1.240 ± 0.012 1.917 ± 0.048 2.453 ± 0.056

MW (g/mol)

843.15 ± 9.52 875.17 ± 10.28 864.04 ± 9.21 855.50 ± 3.69 857.82 ± 4.01

Viscosity (mm2/s) @ 60 °C

14.902 ± 0.193 17.069 ± 0.137 17.122 ± 0.123 17.717 ± 0.150 19.185 ± 0.392

The properties of the mixture M1 (i.e., 100 % RPO) are similar to those reported by Kansedo

et al. (2009) and by Singh and Dipti (2010). Concerning the waste cooking oils properties, they

are quite dependent of the vegetable oil feedstocks and their frying practices and conditions.

The WCO (M5) used in this work has properties similar to those reported by Wan Omar et al.

(2011b) and Lam et al. (2010) and it can be categorized as yellow grease (FFA < 15 %) (Avhad

and Marchetti 2015).

Regarding the mixtures prepared with RPO and WCO, one observes that the density and the

molecular weight are not affected by the blending ratio. On the other hand, the properties


29

related to acidity of the mixtures (𝐴𝑉 and FFA) rise significantly as the percentage of WCO

increases in the blend. The water content and the viscosity are properties that increase slightly

by increasing the WCO percentage in the blend.

2.3.2 Catalysts characterization

The solid catalysts prepared by the methods shown in Table 2.1 were characterized in terms

of some textural properties such as surface area, crystalline structure, but also their basic and

acid strength, etc. The results are shown and discussed below.

2.3.2.1 BET surface area and Hammett indicators analyses

The BET surface area, pore volume, pore diameter, basic and acid strength of catalysts are

shown in Table 2.3.

Table 2.3 - Textural properties of the catalysts prepared in this work.

Catalyst

Specific surface

area (m2/g)

Pore volume (cm3/g)

Pore diameter (Å)

Basic strength Acid strength

FAD 9.0280 0.01055 77.188 9.3≤ pKa <12.2 6.8≤ pKa <7.2

FAC 5.1750 0.00791 101.849 9.3≤ pKa <12.2 6.8≤ pKa <7.2

Dolomite C 12.0113 0.04145 136.908 12.2≤ pKa <15 6.8≤ pKa <7.2

Dolomite CSC 15.2617 0.05291 113.689 9.3≤ pKa <12.2 6.1≤ pKa <6.8

CaO-SiO2 6.6112 0.01285 56.874 7.2≤ pKa <9.3 3.8≤ pKa <4.7

CaO-S-SiO2 12.6773 0.04330 109.925 9.3≤ pKa <12.2 6.8≤ pKa <7.2

CA-PET 1105.2 0.85871 14.983 ND 6.1≤ pKa <6.8

CA-PET-S 624.3 0.54221 14.871 ND 3.8≤ pKa <4.7

ND – not detected

The calcination of fly ashes seems to reduce the surface area (ca 40 %) and pore volume

(ca 25 %), which can be due to sintering of the compounds on the solid matrix surface


30

(Maneerung et al. 2015). Nevertheless, this thermal treatment does not affect both basic and

acid strength of this catalyst. Both ash based catalysts have an intermediate basic strength

(9.3≤ pKa <12.2) and a low acid strength (6.8≤ pKa <7.2).

The sulfonation of Dolomite increases both surface area (ca 30 %) and pore volume (ca 30 %),

but reduces the pore diameter (ca 20 %). These physical changes may have effects on the

performance of these materials in the catalysis of FAME production reactions. On one hand,

higher surface area and pore volume will have a positive effect on the catalysis, but on the

other hand, a decrease of pore diameter increases the diffusion limitations especially for

molecules having long alkyl chain (Lam et al. 2009). Jacobson et al. (2008) identified the pore

structure as the primary requirement for an ideal solid catalyst in the biodiesel production (via

transesterification) since a typical triglyceride molecule has a diameter of approximately 58 Å.

As foreseen, sulfonating the Dolomite C increases its acid strength (6.8≤ pKa <7.2 to 6.1≤ pKa

<6.8) and decreases the basic strength (12.2≤ pKa <15 to 9.3≤ pKa <12.2), values close to those

found by Yoosuk et al. (2011). However, Dolomite CSC has both acid and basic strength which,

a priori, gives it a bifunctional character.

The sulfonation of CaO-SiO2 catalyst enhances considerably the three textural properties:

surface area, pore volume and pore diameter but decreases the acid strength (3.8≤ pKa <4.7

to 6.8≤ pKa <7.2) and increases the basic strength (7.2≤ pKa <9.3 to 9.3≤ pKa <12.2). The effect

of sulfonation and calcination on these strengths could be due to the formation of new phases

of basic character such as calcium sulfate and calcium silicate (Chen et al. 2015b).

In regard to the catalysts prepared from PET both have an acid character being CA-PET-S the

strongest (3.8≤ pKa <4.7), and the basic strength was not detected in none. The acid character

of this carbon catalyst may promote the esterification reaction of FFA but not the

transesterifcation of triglycerides (Borges and Díaz 2012; Fadhil et al. 2016). The sulfonation

treatment performed on the PET catalyst reduced its specific surface area (ca 44%) and pore

volume (ca 37 %), which should be ascribed to the modification of a large number

of – SO3H groups in the carbon framework.


31

In short, all catalysts prepared in this work could be classified as mesoporous catalysts since

the pore diameters are within the intermediate range (20 – 500 Å) (Corma 1997), except for

the carbonaceous catalysts (PET) which is microporous (< 20 Å). This feature may influence the

catalysts’ performance in the transesterification reaction since, as stated before, a typical

triglyceride molecule has a diameter of around 58 Å. Hence, as larger are the porous the higher

is the accessibility of those molecules to the inner pore structure network.

2.3.2.2 SEM and EDX analyses

The SEM images for characterizing the morphological characteristics and EDX for elemental

analysis or chemical characterization of the catalysts were obtained. Figure 2.1 shows the

morphological and the elemental composition of FAD and FAC catalysts. All particles of both

ash catalysts have uniform distribution of agglomerates with irregular shapes, and the

morphological sizes of the particles were reduced by the calcination treatment (Figure 2.1

a&e), possibly due to sintering processes, which decreases the surface area (Ho et al. 2014;

Muthukumaran et al. 2015). The results of EDX show as predominant elements in these

catalysts: Ca, Mg, Si, Al, O, K, S, Na, Cl and P. These elements remained on the solid surface

after calcination, as shown in Figure 2.1 b&c and Figure 2.1 d&f for FAD and FAC, respectively.


32

(a) (b)

(c) (d)

(e) (f)

Figure 2.1 - FAD catalyst: SEM (a) and EDX (b and c); FAC catalyst: SEM (e) and EDX (d and f).


33

Figure 2.2 displays the morphology of solid catalysts Dolomite C (Figure 2.2 a) and Dolomite

CSC (Figure 2.2 b&c). The Dolomite C has a dense surface with heterogeneous distribution of

particle sizes (i.e., irregular size) and smooth appearance, which should be derived from

decarbonation process (calcination) of dolomite rock (Correia et al. 2015; Yoosuk et al. 2011).

Sulfonation caused the elongation of the crystalline structures as fibers due to sulfur

compounds formation, as depicted in Figure 2.2 b&c.

(a) (b) (c)

Figure 2.2 - Dolomite C: SEM (a); Dolomite CSC: SEM (b) and EDX (c).

The catalyst prepared from eggshells CaO-SiO2 exhibits large and regular blocks particles

(Figure 2.3 a). The same was observed in the CaO-S-SiO2 catalyst (image not shown) and one

infers that this morphology could be owed to the coverage of Si compounds on the CaO surface

(see Figure 2.3 b&c). More, the same effect was observed by Chen et al. (2015b). As in the

dolomitic catalysts, the sulfonation of CaO-SiO2 solid also originated the formation of

crystalline structures as flat elongated fibers (Figure 2.3 c) in this eggshells based material. This

phenomenon was also observed by Embong et al. (2016).

(a) (b) (c)

Figure 2.3 - CaO-SiO2: SEM (a): CaO-S-SiO2: EDX (b) and (c).


34

The images taken for catalysts prepared from PET are shown in Figure 2.4, where it can be

seen irregular and flat surface with crevices.

(a) (b)

Figure 2.4 - SEM images of CA-PET (a) and CA-PET-S (b) catalysts.

The sulfonation does not generate observable significant differences in the morphology of this

material, since particles have similar shapes in both photos (a) and (b) of Figure 2.4. Similar

behavior was observed in other studies (Dawodu et al. 2014; Fadhil et al. 2016; Hajamini et al.

2016).

2.3.2.3 XRD analyses

The XRD diffractograms of the fly ash catalysts are depicted in Figure 2.5 a. The structure and

crystalline compounds of FAD and FAC are similar, being only their main differences the area

and the intensity of the peaks after calcination. The XRD pattern for FAD catalyst shows clear

diffraction peaks corresponding to calcium oxide (CaO) phase detected at 2=32.2o, 37.4o,

53.8o, 65.2o, and 67.5o, calcium carbonate (CaCO3-major component) phase detected at

2=23.3o, 29.6o, 36.2o, 39.7o, 43.4o, 47.8o, 48.8o, 56.9o, 61.0o and 65.0o, potassium chloride

(KCl) phase detected at 2=28.5o, 40.5o, and silicon dioxide (SiO2) phase detected at 2 =20.9o,

26.7o, 36.38o, 39.46o, 40.28o, 50.2o, 60.2o and 68.5o, among other components. After the ash

calcination, i.e. for FAC catalyst, CaCO3 was transformed into CaO (Chen et al. 2015a; Sharma


35

et al. 2012) and this is evident by the higher intensity of the corresponding peak. This latter is

the major component in FAC followed by the silicon dioxide (SiO2).

Figure 2.5 b shows the XRD of Dolomite C and Dolomite CSC catalysts. The presence of both

phases: CaO (2θ=32.2o) and MgO (2θ=42.7o) in the Dolomite C could promote the

transesterification reaction. Ca(OH)2 (2=34.1o) is part of the chemical composition of these

catalysts, its formation occurs readily upon an exposure of CaO to humidity of ambient,

resulting in a significant loss of the transesterification activity (Correia et al. 2015; Jaiyen et al.

2014). It seems that calcination time of dolomite rock was sufficient to decompose MgCO3 in

to MgO, but not enough to convert completely the CaCO3 in to CaO, since CaCO3 is present

(2=29.3o) in this catalyst after that thermal treatment; similar result was observed by

Ngamcharussrivichai et al. (2010). The sulfonation of Dolomite C originated new peaks in the

diffractogram (of Dolomite CSC), corresponding to calcium sulfate (CaSO4-major component)

at 2=25.6o, 31.3o, 38.6o, 40.9o, 48.6o, 52.2o, 55.8o, and 65.0o. As discussed previously, this

treatment had also effects on the basic and acid strengths of the solid catalyst due to the

replacement of calcium carbonate by calcium sulfate, which in turn could affect its catalytic

activity.


36

(a) (b)

(c) (d)

Figure 2.5 - XRD patterns of catalysts: FAD and FAC (a), Dolomite C and Dolomite CSC (b), CaO-SiO2

and CaO-S-SiO2 (c), and CA-PET and CA-PET-S (d). (• SiO2, ■ CaCO3, ♦ CaO, ▲ CaSO4, ■

CaCO3, ▼ Ca(OH)2, ♦ CaO, □ Ca2SiO4 and ★ MgO).

The XRD patterns of the catalysts produced from eggshells are shown in Figure 2.5 c. For CaO-

SiO2 catalyst the peaks at 2= 37.2o, 64.2o, 76.1o and 2= 20.2o, 33.4o, 39.7o, 55.3o, 59.8o

correspond to CaO and Ca(OH)2, respectively. Besides, calcium silicate compounds (Ca2SiO4)

peaks appear at 2 = 23.3o, 26.2o, 28.0o 32.9o, 35.1o, 41.2o due to the reaction of Na2SiO3 with

CaO and Ca(OH)2 during the catalyst preparation process (Leite et al. 2017). For the CaO-S-SiO2

catalyst new peaks have arose at 2 = 25.6o, 31.3o, 48.6o and at 2 = 29.3o, corresponding to

CaSO4 and CaCO3, respectively (Nur Syazwani et al. 2017). There are the three main

compounds that can be identified in the CaO-SiO2 catalyst, namely: CaO, Ca(OH)2 and Ca2SiO4.

The latter two are the most abundant, which means that part of the calcium existing in the

eggshells has reacted with Na2SiO3 to form calcium silicate (Leite et al. 2017). After CaO-SiO2


37

sulfonation, the hydroxides and silicates of calcium were mainly converted into CaSO4, the

predominant compound in this catalyst.

Finally, the XRD diffractograms of catalysts prepared from PET are depicted in Figure 2.5 d.

Both diffractograms exhibit a broad diffraction peaks indicating an amorphous carbons, C(002)

and C(101), which is composed of the oriented random fashion of carbon sheets. Therefore,

both samples are composed of high non-graphitic carbon content. Fadhil et al. (2016) and

Chang et al. (2015) observed that these kind of carbons have oriented random fashion sheets.

In short, both catalysts prepared from PET have high content of non-graphitic carbon.

2.3.2.4 FTIR analyses

The FTIR spectra of all catalysts prepared in this work are shown in Figure 2.6. The FTIR

spectrum of FAD (Figure 2.6 a) shows the major absorption broad band at 1408.1 cm-1 and

minor absorption bands at 875.5 and 711.2 cm-1, which correspond to the asymmetric

stretching and to out-of-plane band and in-plane band vibration modes of carbonate (CO3-2)

group, respectively. This result confirms the presence of CaCO3 in FAD, detected by XRD.

PO4-3 and Si-O components (silica phosphates) show broad bands in the region between 1100.5

and 911.6 cm-1; the same was observed by Maneerung et al. (2015) and Sharma et al. (2012)

but using bottom ash waste arising from woody biomass gasification and wood ash from the

Acacia nilotica (babul), respectively. Moreover, the absorption sharp band at 3643 cm-1, which

is attributed to -OH band, was observed for both catalysts (calcined and uncalcined). This band

is an evidence of water absorption on the CaO surface producing Ca(OH)2 (Boey et al. 2011).

The typical transmittance FTIR spectra of the Dolomite C and Dolomite CSC are shown in Figure

2.6 b. The bands at 1442.2 and 1438 cm-1 can be assigned to the symmetric and asymmetric

stretching vibrations of O–C–O bonds of unidentate carbonate at the surface of the calcium–

magnesium oxide in both dolomitic catalysts (Algoufi et al. 2017). The band at 872.2 cm-1 arises

also from these carbonates groups. For the Dolomite CSC, the peaks at 1098.9, 672.6, 611.2


38

and 593.8 cm-1 are attributed to the functional group SO4-2 of calcium sulfate (major

component) (Kong et al. 2012; Zhao et al. 2016).

(a)

(b)

Figure 2.6 - FTIR spectra of catalysts: FAD and FAC (a), Dolomite C and Dolomite CSC (b), CaO-

SiO2 and CaO-S-SiO2 (c), and CA-PET and CA-PET-S (d).


39

(c)

(d)

Figure 2.6 (cont.)- FTIR spectra of catalysts: FAD and FAC (a), Dolomite C and Dolomite CSC

(b), CaO-SiO2 and CaO-S-SiO2 (c), and CA-PET and CA-PET-S (d).


40

Regarding the FTIR spectra of CaO-SiO2 and CaO-S-SiO2 catalysts (Figure 2.6 c), one observes

nearby absorption bands, such as at 3641.9 and 3653.8 cm-1 that correspond to the stretching

O-H due to physisorption of water on the solid surface. The spectra show matching bands

namely at 937.5 and 938.9 cm-1, which belong to Si-O symmetric elongation vibrations and

Si–O–Ca (Boro et al. 2011; Chen et al. 2015a). The absorption band at 1128.6 cm-1 could be

attributed to O–Si–O bond of silicate compounds (Amani et al. 2016). For CaO-S-SiO2 catalyst,

the bands at 1095.7, 672.6, 610.3 and 592.8 cm-1 are attributed to the stretching vibrations of

S=O on the group SO4-2 of calcium sulfate (Kong et al. 2012; Zhao et al. 2016).

Concerning the FTIR spectra of CA-PET and CA-PET-S catalysts (Figure 2.6 d), both are similar

in respect to the position of their bands. The absorption bands observed at 1008, 1030 and

1120 cm-1 are assigned to the symmetric stretching vibrations of S=O as result of inducing the

SO3H group (Chang et al. 2015; Xing et al. 2007). These evidences indicate a successful

incorporation of SO3H functional groups in the carbon framework.

2.3.2.5 Catalysts performance

The performance of the catalysts prepared from waste materials was assessed through the

esterification and transesterification reaction yields, in the conversion of RPO and WCO

mixtures to FAME. The results are plotted in Figure 2.7 for FAME yield (Equation 2.4) and FFA

conversion (Equation 2.5). As FAME are produced by both the esterification and the

transesterification reaction, the FAME yield reflects the global conversion of these two

reactions. However, in this work, the FFA content is low, so the contribution of the

esterification reaction to the FAME yield will be much less than that of the transesterification

reaction.

Regarding the results obtained in this work, as the WCO percentage in the reaction mixture

increases two main conclusions are withdrawn by an overview of Figure 2.7, namely:


41

A decrease of FAME yield for catalysts with moderate or high basic strength. This can

be due to the neutralization of their basic catalyst sites, as stated by Kaur et al. (2011)

and Kouzu et al. (2008);

An increase (or maintenance, for FAD) of FFA conversion.

Thus, the presence of FFA in the reaction mixture favors their conversion to FAME but has a

negative effect on transesterification reaction yield.

A more detailed analysis of each catalyst group shows that the fly ash catalysts (Figure 2.7 a&b)

promoted the highest FAME yield for all oil blends tested. One of the best performances (of all

catalysts) in FFA conversion is also related to one of these catalysts, the FAD, achieving values

above 96 % for all mixtures. However, rising the WCO in the reaction mixture decreases the

FAME yield in ca. 30 to 35 %, being the FAC performance the most affected.

Among the two ash based catalysts, the FAC is the one that has a worse performance in both

FAME yield and FFA conversion. This evidence may be due to changes in surface morphology

(sintering processes) and decrease of both crystalline phases and active functional groups on

the surface for the calcined catalyst, as observed by XRD and FTIR. Additionally, the most

abundant compound in FAD is CaCO3 and in FAC is CaO, which may be another reason for the

observed differences in performance. However, as the amount of WCO in the blend increases

these differences in performance of the catalyst decrease significantly.

Despite of the low acid strength of both catalysts (6.8≤ pKa <7.2), the FAD roughly converts all

FFA present in the initial reaction mixture and the FAC increases its performance as the WCO

percentage rises, achieving ca. 86 % of conversion for 100 wt% of WCO (M5). Summing up, it

is reasonable to conclude that these catalysts have a bifunctional character, especially the FAD

which enables attaining FAME yield and FFA conversions around 95 % for WCO blends of

25 wt% (M2).


42

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 2.7 - Performance in terms of FAME yield and FFA conversion of catalysts: FAD and FAC

(a) and (b), Dolomite C and Dolomite CSC (c) and (d), CaO-SiO2 and CaO-S-SiO2 (e)

and (f), and CA-PET and CA-PET-S(g) and (h), for several RPO:WCO mixtures.


43

The FAME yields and FFA conversions attained using Dolomite catalysts are summarized in the

Figure 2.7 c and Figure 2.7 d, respectively. Dolomite C has the highest basic strength

(12.2≤ pKa <15), but this is not reflected in higher FAME yields. Furthermore, in comparison to

ash catalysts, Dolomite C has higher surface area, pore volume and pore diameter, and all

these features do not seem to be sufficient to give a better performance to this catalyst. Thus,

chemical composition of the catalyst could have a stronger influence than those physical

characteristics. The highest FAME yield achieved by this catalyst is ca. 88 % for M1, and

decreases 37 % for the mixture with higher acid value (M5). With regard to FFA conversion,

the values attained are consistent with the low acid strength exhibited by this catalyst, i.e.,

Dolomite C has a low acid strength and consequently is a weak catalyst of esterification

reaction, so the FFA conversion values reached are one of the lowest registered (for each oil

mixture).

The sulfonation and subsequent calcination of Dolomite C, giving rise to Dolomite CSC, has

strongly affected its ability to catalyze the transesterification reaction, nearly annulling it. This

could be due to the decrease of basic strength (by neutralization) and/or to the decrease of

pore diameter, which in turn increases the diffusion limitations for long alkyl chain molecules

(e.g., triglycerides). However, this treatment has improved the performance of this material

for catalyzing the esterification reaction, reaching conversions of 100 % for oil mixtures with

the highest acid values (M4 and M5) tested in this work. The moderate acid character of this

catalyst (6.1≤ pKa <6.8) and the functional groups on its surface (group SO4-2) could be the

driving force of its performance in converting FFA to FAME.

The data concerning the performance of eggshell catalysts (CaO-SiO2 and CaO-S-SiO2) are

plotted in the Figure 2.7 e&f. The FAME yield are negatively affected by the initial acid value

of the oil mixtures, achieving the lowest values (of all catalysts) for M5. The highest yield levels

reached was 70 % for M1 in CaO-S-SiO2 catalyst and 40 % for M2 in CaO-SiO2; similar results

were obtained by Chen et al. (2015b). The sulfonation and calcination of Ca-SiO2 material

increases the FAME yield, possibly due to the formation of new active phases such as calcium


44

sulfate (CaSO4) and calcium carbonate formed from calcium hydroxide (Ca(OH)2) and calcium

silicate (Ca2SiO4), as shown in the Figure 2.5 c, which have low activity towards the

transesterification reaction. Relating to the performance of eggshells based catalysts to

converting the FFA to FAME, one can say that as WCO percentage in the oil mixture increases

the higher is the conversion, being the Ca-SiO2 catalyst the best one. This finding is in

agreement with the acid strength of that catalyst (Ca-SiO2), which is one of the highest

(3.8≤ pKa <4.7) observed in this work.

Compared to the several catalysts discussed above, those produced from PET exhibit an

inverse trend over the FAME yield (Figure 2.7 g). It was registered an increase of yield as the

amount of FFA in the oil mixtures rises, but the higher values attained were low, i.e., ca. 25 %

and 39 % for M5 in CA-PET and CA-PET-S, respectively. This result can be explained considering

the lack of basic strength in these catalysts (see Table 2.3). In fact, PET catalysts only have an

acid character, being CA-PET-S the catalyst with the most acidic character produced in this

work (3.8 ≤pKa <4.7). Though, they do not have the best performance in the esterification

reaction catalysis. In Figure 2.7 h one observes that the FFA conversions increases as the acid

value of oil mixture rises, achieving the maximum values of ca.71 % and 87 % for M5 in CA-PET

and CA-PET-S, respectively. Thus, although these catalysts have the largest specific surface

area, the largest pore volume and one of the highest acidic strengths, this does not seem to

be enough to make them the best catalysts for the esterification reaction. A plausible reason

for this may lie in the fact that they have the smallest pore size observed among the catalyst

developed in this work. In addition, among the two PET catalysts, CA-PET-S is the one that has

a better performance in the catalysis of both esterification and transesterification reactions.


45

2.4 Conclusions

In the present study, efficient heterogeneous catalysts were successfully prepared from solid

waste materials for biodiesel production by transesterification and esterification, using

mixtures of refined palm oil and waste cooking oil in different ratios and methanol. The results

demonstrate that all the catalysts evaluated have different catalytic performances. The better

catalyst for catalyzing simultaneous both transesterification and esterification reactions, i.e.,

having a bifunctional character, was biomass fly ash dried (FAD), achieving yields and

conversions above 95 % for a blend of up to 25 wt% of WCO.

The catalyst produced from dolomite rock did not have a strong bifunctional character. They

showed good performances in catalyzing the transesterification and esterification reactions,

but not simultaneously. Indeed, the sulfonation of Dolomite C was aimed at increasing its

acidic strength so as to give it the potential to catalyze the esterification reaction. However,

that treatment strongly affected that ability, practically canceling it. In further procedures, the

sulfonation stage should be more lenient.

The sulfonation of material prepared from eggshells improved its ability for catalyzing the

transesterification reaction, however the maximum values attained do not exceed 70 % of

yields. On the other hand, this treatment worsened the performance of this catalyst in FFA

conversion.

Regarding the catalysts produced from PET, the results showed that they are good candidates

for catalyzing the esterification reaction of high acid value feedstocks.

In short, none of the catalysts produced in this work has both high basic and acidic strengths

and the only one that has these two strengths balanced (on a moderate level) is the Dolomite

CSC. However, the catalyst that exhibited a bifunctional character was undoubtedly FAD, which

means this material can be directly and immediately used from the electrostatic precipitator


46

equipment located at the biomass thermal power-plant, as its moisture content is very low,

with subsequent economic benefits.

By the exploitation of residual feedstocks (e.g., WCO) and the use of waste based catalysts,

this work gives a contribution to make the biodiesel production a low cost, affordable and

sustainable process, and simultaneously minimizing the environmental burdens traditionally

inherent to the management of those wastes. Therefore, an awareness should be created so

that any material that is deemed to be waste could be exploited for usage in this or other

applications, thereby implementing the principles of circular economy.

Acknowledgments

Edgar M. Vargas S. express his sincere gratitude to the Jorge Tadeo Lozano University of

Colombia (Direction of Investigation, Creation and Extension) for the financial assistance of

this work. The authors thanks for the financial support to CESAM (UID/AMB/50017 - POCI-01-

0145-FEDER-007638), funded by national funds (FCT/MCTES) through PIDDAC and co-funded

by the FEDER, within the PT2020 Partnership Agreement and Compete 2020.


47

References

AENOR-EN 14103. 2011. Productos Derivados de Aceites y Grasas. Ésteres Metílicos de Ácidos Grasos (FAME). Determinación de Los Contenidos de Éster y de Éster Metílico Del Ácido Linolénico. España. www.agilent.com/chem. (February 15, 2019).

Algoufi, Y. T., G. Kabir, and B. H. Hameed. 2017. “Synthesis of Glycerol Carbonate from Biodiesel By-Product Glycerol over Calcined Dolomite.” Journal of the Taiwan Institute of Chemical Engineers 70: 179–87. http://dx.doi.org/10.1016/j.jtice.2016.10.039.

Amani, H., M. Asif, and B. H. Hameed. 2016. “Transesterification of Waste Cooking Palm Oil and Palm Oil to Fatty Acid Methyl Ester Using Cesium-Modified Silica Catalyst.” Journal of the Taiwan Institute of Chemical Engineers 58: 226–34.

ASTM. 2010. “Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity).” Annual Book of ASTM Standards: 1–10.


Boey, Peng-Lim, Gaanty Pragas Maniam, and Shafida Abd Hamid. 2009. “Biodiesel Production via Transesterification of Palm Olein Using Waste Mud Crab (Scylla Serrata) Shell as a Heterogeneous Catalyst.” Bioresource technology 100(24): 6362–68. http://www.ncbi.nlm.nih.gov/pubmed/19666218 (May 5, 2015).

Boey, Peng Lim et al. 2011. “Utilization of BA (Boiler Ash) as Catalyst for Transesterification of Palm Olein.” Energy 36(10): 5791–96. http://dx.doi.org/10.1016/j.energy.2011.09.005.

Borges, M.E., and L. Díaz. 2012. “Recent Developments on Heterogeneous Catalysts for Biodiesel Production by Oil Esterification and Transesterification Reactions: A Review.” Renewable and Sustainable Energy Reviews 16(5): 2839–49. http://linkinghub.elsevier.com/retrieve/pii/S1364032112000834 (July 11, 2014).

Boro, Jutika, Ashim J. Thakur, and Dhanapati Deka. 2011. “Solid Oxide Derived from Waste Shells of Turbonilla Striatula as a Renewable Catalyst for Biodiesel Production.” Fuel Processing Technology 92(10): 2061–67. http://dx.doi.org/10.1016/j.fuproc.2011.06.008.

Chang, Binbin et al. 2015. “Synthesis of Sulfonated Porous Carbon Nanospheres Solid Acid by a Facile Chemical Activation Route.” Journal of Solid State Chemistry 221: 384–90. http://dx.doi.org/10.1016/j.jssc.2014.10.029.


48

Chen, Guan Yi, Rui Shan, Jia Fu Shi, and Bei Bei Yan. 2015a. “Transesterification of Palm Oil to Biodiesel Using Rice Husk Ash-Based Catalysts.” Fuel Processing Technology 133: 8–13. http://dx.doi.org/10.1016/j.fuproc.2015.01.005.

Chen, Guanyi, Rui Shan, Shangyao Li, and Jiafu Shi. 2015b. “A Biomimetic Silicification Approach to Synthesize CaO-SiO2 Catalyst for the Transesterification of Palm Oil into Biodiesel.” Fuel 153: 48–55. http://dx.doi.org/10.1016/j.fuel.2015.02.109.

Chen, Jein Wen et al. 2012. “Carcinogenic Potencies of Polycyclic Aromatic Hydrocarbons for Back-Door Neighbors of Restaurants with Cooking Emissions.” Science of the Total Environment 417–418: 68–75. http://dx.doi.org/10.1016/j.scitotenv.2011.12.012.

Corma, Avelino. 1997. “From Microporous to Mesoporous Molecular Sieve Materials and Their Use in Catalysis.” Chemical Reviews 97(6): 2373–2420. http://pubs.acs.org/doi/abs/10.1021/cr960406n.

Correia, Leandro Marques et al. 2015. “Characterization and Application of Dolomite as Catalytic Precursor for Canola and Sunflower Oils for Biodiesel Production.” Chemical Engineering Journal 269: 35–43. http://dx.doi.org/10.1016/j.cej.2015.01.097.

Dawodu, Folasegun A. et al. 2014. “Effective Conversion of Non-Edible Oil with High Free Fatty Acid into Biodiesel by Sulphonated Carbon Catalyst.” Applied Energy 114: 819–26. http://dx.doi.org/10.1016/j.apenergy.2013.10.004.


Embong, Nurul Hajar et al. 2016. “Utilization of Palm Fatty Acid Distillate in Methyl Esters Preparation Using so 42-/TiO2-SiO2as a Solid Acid Catalyst.” Journal of Cleaner Production 116: 244–48. http://dx.doi.org/10.1016/j.jclepro.2015.12.108.

Fadhil, Abdelrahman B., Akram M. Aziz, and Marwa H. Al-Tamer. 2016. “Biodiesel Production from Silybum Marianum L. Seed Oil with High FFA Content Using Sulfonated Carbon Catalyst for Esterification and Base Catalyst for Transesterification.” Energy Conversion and Management 108: 255–65.

Farooq, Muhammad, Anita Ramli, and Abdul Naeem. 2015. “Biodiesel Production from Low FFA Waste Cooking Oil Using Heterogeneous Catalyst Derived from Chicken Bones.” Renewable Energy 76: 362–68. http://linkinghub.elsevier.com/retrieve/pii/S0960148114007575 (February 6, 2015).


49

Farooq, Muhammad, Anita Ramli, and Duvvuri Subbarao. 2013. “Biodiesel Production from Waste Cooking Oil Using Bifunctional Heterogeneous Solid Catalysts.” Journal of Cleaner Production 59: 131–40. http://linkinghub.elsevier.com/retrieve/pii/S095965261300396X (May 14, 2015).

Hajamini, Zahra, Mohammad Amin Sobati, Shahrokh Shahhosseini, and Barat Ghobadian. 2016. “Waste Fish Oil (WFO) Esterification Catalyzed by Sulfonated Activated Carbon under Ultrasound Irradiation.” Applied Thermal Engineering 94: 1–10. http://dx.doi.org/10.1016/j.applthermaleng.2015.10.101.


ICONTEC. 2011. NTC 218. Grasas y Aceites Vegetales y Animales. Determinación Del Índice de Acidez y de La Acidez. Colombia. https://tienda.icontec.org/wp-content/uploads/pdfs/NTC218.pdf.

ICONTEC. 1999. NTC 218. “Grasas Y Aceites Vegetales Y Animales. Determinacion de Índice de Acidez.” (571). Colombia. https://tienda.icontec.org/wp-content/uploads/pdfs/NTC218.pdf.

Jacobson, Kathlene, Rajesh Gopinath, Lekha Charan Meher, and Ajay Kumar Dalai. 2008. “Solid Acid Catalyzed Biodiesel Production from Waste Cooking Oil.” Applied Catalysis B: Environmental 85(1–2): 86–91.

Jaiyen, Siyada, Thikumporn Naree, and Chawalit Ngamcharussrivichai. 2014. “Comparative Study of Natural Dolomitic Rock and Waste Mixed Seashells as Heterogeneous Catalysts for the Methanolysis of Palm Oil to Biodiesel.” Renewable Energy 74: 433–40. http://dx.doi.org/10.1016/j.renene.2014.08.050.

Kansedo, Jibrail, Keat Teong Lee, and Subhash Bhatia. 2009. “Cerbera Odollam (Sea Mango) Oil as a Promising Non-Edible Feedstock for Biodiesel Production.” Fuel 88(6): 1148–50. http://dx.doi.org/10.1016/j.fuel.2008.12.004.

Kaur, Mandeep, and Amjad Ali. 2011. “Lithium Ion Impregnated Calcium Oxide as Nano Catalyst for the Biodiesel Production from Karanja and Jatropha Oils.” Renewable Energy 36(11): 2866–71. http://dx.doi.org/10.1016/j.renene.2011.04.014.


50

Kong, Bao et al. 2012. “Synthesis of α-Calcium Sulfate Hemihydrate Submicron-Rods in Water/n-Hexanol/CTAB Reverse Microemulsion.” Colloids and Surfaces A: Physicochemical and Engineering Aspects 409: 88–93. http://dx.doi.org/10.1016/j.colsurfa.2012.05.041.

Kouzu, Masato et al. 2008. “Calcium Oxide as a Solid Base Catalyst for Transesterification of Soybean Oil and Its Application to Biodiesel Production.” Fuel 87(12): 2798–2806.

Lam, Man Kee, and Keat Teong Lee. 2010. “Accelerating Transesterification Reaction with Biodiesel as Co-Solvent: A Case Study for Solid Acid Sulfated Tin Oxide Catalyst.” Fuel 89(12): 3866–70. http://dx.doi.org/10.1016/j.fuel.2010.07.005.

Lam, Man Kee, Keat Teong Lee, and Abdul Rahman Mohamed. 2009. “Sulfated Tin Oxide as Solid Superacid Catalyst for Transesterification of Waste Cooking Oil: An Optimization Study.” Applied Catalysis B: Environmental 93(1–2): 134–39.

Leite, F. H G, T. F. Almeida, R. T. Faria, and J. N F Holanda. 2017. “Synthesis and Characterization of Calcium Silicate Insulating Material Using Avian Eggshell Waste.” Ceramics International 43(5): 4674–79. http://dx.doi.org/10.1016/j.ceramint.2016.12.146.


Maneerung, Thawatchai, Sibudjing Kawi, and Chi Hwa Wang. 2015. “Biomass Gasification Bottom Ash as a Source of CaO Catalyst for Biodiesel Production via Transesterification of Palm Oil.” Energy Conversion and Management 92: 234–43. http://dx.doi.org/10.1016/j.enconman.2014.12.057.

Mansir, Nasar et al. 2018. “Modified Waste Egg Shell Derived Bifunctional Catalyst for Biodiesel Production from High FFA Waste Cooking Oil. A Review.” Renewable and Sustainable Energy Reviews 82(May 2017): 3645–55.

Muthukumaran, N. et al. 2015. “Synthesis of Cracked Calophyllum Inophyllum Oil Using Fly Ash Catalyst for Diesel Engine Application.” Fuel 155: 68–76. http://dx.doi.org/10.1016/j.fuel.2015.04.014.

Ngamcharussrivichai, Chawalit, Pramwit Nunthasanti, Sithikorn Tanachai, and Kunchana Bunyakiat. 2010. “Biodiesel Production through Transesterification over Natural Calciums.” Fuel Processing Technology 91(11): 1409–15. http://dx.doi.org/10.1016/j.fuproc.2010.05.014.


51

Nur Syazwani, O. et al. 2017. “Esterification of High Free Fatty Acids in Supercritical Methanol Using Sulfated Angel Wing Shells as Catalyst.” Journal of Supercritical Fluids 124: 1–9. http://dx.doi.org/10.1016/j.supflu.2017.01.002.



Singh, S. P., and Dipti Singh. 2010. “Biodiesel Production through the Use of Different Sources and Characterization of Oils and Their Esters as the Substitute of Diesel: A Review.” Renewable and Sustainable Energy Reviews 14(1): 200–216.

Uprety, Bijaya K., Wittavat Chaiwong, Chinomnso Ewelike, and Sudip K. Rakshit. 2016. “Biodiesel Production Using Heterogeneous Catalysts Including Wood Ash and the Importance of Enhancing Byproduct Glycerol Purity.” Energy Conversion and Management 115: 191–99. http://dx.doi.org/10.1016/j.enconman.2016.02.032.

Wan Omar, Wan Nor Nadyaini, and Nor Aishah Saidina Amin. 2011a. “Biodiesel Production from Waste Cooking Oil over Alkaline Modified Zirconia Catalyst.” Fuel Processing Technology 92(12): 2397–2405. http://linkinghub.elsevier.com/retrieve/pii/S0378382011003031 (January 2, 2015).

Wan Omar, Wan Nor Nadyaini, and Nor Aishah Saidina Amin. 2011b. “Optimization of Heterogeneous Biodiesel Production from Waste Cooking Palm Oil via Response Surface Methodology.” Biomass and Bioenergy 35(3): 1329–38. http://linkinghub.elsevier.com/retrieve/pii/S0961953410005076 (May 29, 2015).

Xing, Rong et al. 2007. “Active Solid Acid Catalysts Prepared by Sulfonation of Carbonization-Controlled Mesoporous Carbon Materials.” Microporous and Mesoporous Materials 105(1–2): 41–48.

Yoosuk, Boonyawan, Parncheewa Udomsap, and Buppa Puttasawat. 2011. “Hydration–Dehydration Technique for Property and Activity Improvement of Calcined Natural Dolomite in Heterogeneous Biodiesel Production: Structural Transformation Aspect.” Applied Catalysis A: General 395(1–2): 87–94. http://linkinghub.elsevier.com/retrieve/pii/S0926860X11000421 (May 29, 2015).

Zhao, Yuzeng et al. 2016. “Inhibition of Calcium Sulfate Scale by Poly (Citric Acid).” Desalination 392: 1–7. http://dx.doi.org/10.1016/j.desal.2016.04.010.


52

53

SECTION C - Optimization of FAME production in batch mode operation

Section C presents the optimization study of the FAME production process (in batch mode)

with methanol, using mixtures of WCO and RPO and biomass fly ash as a catalyst (chosen in

section B). The Box–Behnken experimental design and the Response Surface Methodology

were used in the optimization, and four operating variables were tested, namely: catalyst

loading, methanol/oil molar ratio, RPO/WCO mass ratio and reaction temperature.

Furthermore, the catalyst was characterized by SEM, EDX, XRD, BET, FT-IR and Hammett

indicators. Additionally, it was carried out a study of reusability of the catalyst aiming to assess

its performance over several cycles of utilization in the FAME synthesis process. The optimal

operating conditions found by the regression model were used in this set of assays and the

catalyst was characterized after each reuse by XRD, BET and Hammett indicators.

The information presented in this section was adapted from the following published article:

E. M. Vargas, J. L. Ospina, M. C. Neves, L. A. C. Tarelho, and M. I. Nunes, “Optimization

of FAME production from blends of waste cooking oil and refined palm oil using

biomass fly ash as a catalyst” Renew. Energy, vol. 163, pp. 1637-1647, 2020, doi:

10.1016/j.renene.2020.10.030.


residual materials: Optimization of the process by Box–Behnken model,” Energy

Reports, vol. 6, pp. 347–352, 2019, doi: 10.1016/j.egyr.2019.08.071.


54

OPTIMIZATION OF FAME PRODUCTION IN BATCH OPERATION MODE

55

3 Optimization of FAME production from blends of waste cooking

oil and refined palm oil using biomass fly ash as a catalyst

Abstract: One of the problems associated with biomass combustion is the amount of fly ashes

generated and its subsequent management. The search for ways of valorizing these ashes has

been a challenge for the academic and industrial community. On the other hand, used cooking

oils are wastes which management is quite difficult, by they have a very important energetic

potential. The goal of this work was to optimize the Fatty Acid Methyl Esters (FAME) process,

recovering two residual materials (waste cooking oils (WCO), and biomass fly flash (FAD)). The

optimization of the process was achieved using the response surface methodology and a Box-

Benhken experimental design applied to mixtures of WCO and refined palm oil (RPO), using

FAD as catalyst. The influence on FAME yield of four variables (catalyst loading, methanol/oil

molar ratio, RPO/WCO ratio and reaction temperature) was studied. The higher FAME yield

achieved was 73.8 % for the following operating conditions: 13.57 wt% of catalyst loading, 6.7

of methanol/oil molar ratio, 28.04 wt% of RPO in the oil mixture with WCO and 55 oC for the

reaction temperature. The reusability of the FAD catalyst in the process was also studied

through three successive usage cycles finding no loss of catalytic activity.

Keywords: Biomass fly ash; FAME; optimization; refined palm oil; response surface

methodology; waste cooking oil.


56

3.1 Introduction

The production of biodiesel has become a very important area of research due to the rapid

depletion of energy reserves and the increase in oil prices along with environmental concerns

(Liu et al. 2014). In the current situation, the foremost amount of energy is supplied by the

conventional fossil fuel resources, such as gasoline, liquefied petroleum gas, diesel fuel, coal

and natural gas. It is imperative to find alternative fuels to the petroleum based ones in order

to, along with environmental issues, prolong the petroleum supply. One of the most promising

biofuel is biodiesel, a “green fuel” alternative to diesel fuel, derived from renewable sources

with high quality (Leung et al. 2010). The integration of wastes as a catalyst or as an (vegetable)

oil feedstock into the biodiesel production process can be a promising way to reduce

environmental burdens and the production costs, while also aligning with the principles of

circular economy.



et al. 2018), for the total production costs. The waste cooking oils (WCO) are edible vegetable

oils that have been previously used for frying or cooking and can constitute an additional

source of raw material for biodiesel production. This feedstock can be two to three times

cheaper than virgin vegetable oils (Demirbas 2009; Nurfitri et al. 2013). Furthermore, it is

generally accepted that reusing WCO for human consumption is harmful to health and this

waste is difficult to manage (Chen et al. 2012).

It is important to mention that the catalyst commonly used in the biodiesel production is the

sodium or potassium hydroxide, which have been economically unfeasible to recover from the

process. Aiming to tackle this hotspot of the process, some research (Boey et al. 2009;

Chakraborty et al. 2010; Mendonça et al. 2019) have been focused on the exploitation of waste

materials (e.g. shells, ashes, peels and bones), due to their abundance and low cost, for solid

catalysts preparation. On the other hand, biomass fly ashes (FAD) are a residual materials

whose disposal and management represent a significant challenge (Kotwal et al. 2009), given


57

its increasing production over the last two decades (Girón et al. 2015). The development of

alternative solutions for FAD proper utilization/valorization with emphasis on finding new

applications is currently a very important issue (Chatterjee et al. 2018). The use of FAD, as a

(heterogeneous) catalyst, on the biodiesel production process has been proving to be a

promising alternative to valorize this waste; it has been found that FAD have a potential for

catalyzing the reactions for FAME (Fatty Acid Methyl Esters) production and have bifunctional

characteristics (acid and basic) that allow catalyzing transesterification and esterification

reactions simultaneously (Vargas et al. 2019).

The main objective of this work was to optimize the FAME production process using mixtures

of WCO and refined palm oil, and FAD as catalyst. The effect on FAME yield of four operating

variables was tested, namely: catalyst loading, methanol/oil molar ratio, RPO/WCO mass ratio

and reaction temperature. Additionally, it was carried out a study of reusability of the catalyst

aiming to assess its performance over several cycles of utilization in the FAME synthesis

process.


The Box-Benhken experimental design and Response Surface Methodology (RSM) were used

to design the experiments to optimize the FAME production process and for the data

treatment.

Previously, the solid catalyst (FAD) was prepared and characterized in terms of some of its

chemical, physical and structural properties. The raw-material for FAME synthesis consisted of

a blends of WCO and RPO in different ratios. The adopted procedures are described in the next

sections as well as the analytical methods used.


58

3.2.1 Materials

Waste cooking oil for FAME production was provided by a local collecting company (Bioils) in

Bogotá, Colombia. The WCO was pre-treated by filtration to remove suspended particles and

heating (at 110 °C for 1 h) to eliminate traces of water. The RPO was purchased at a local store

in Bogotá. The FAD came from a dedusting system (electrostatic precipitator) of a thermal

power-plant using residual forest biomass (mainly derived from eucalyptus) sited in the Centre

Region of Portugal.

All the chemicals used were analytical grade except n-hexane (GC grade) and methyl

heptadecanoate (analytical standard) from Sigma-Aldrich and Merck.


According to the experimental design presented in Section 3.2.5, three oily feedstock were

prepared using different mass ratios of RPO/WCO: M1 (100 % RPO), M2 (50 % RPO, 50 % WCO)

and M3 (0 % RPO). These mixtures were characterized in terms of: acid value (NTC

218)(ICONTEC 218 1999), density (NTC 336) (ICONTEC 1998), saponification number (NTC

335)(ICONTEC 1998), viscosity (ASTM D445)(ASTM 2010), and moisture content (Karl Fisher,

Coulometer 831-Metrohm).

The saponification number (SN) was used to calculate the molecular mass (𝑀𝑊, g/mol)

according to Equation 3.1 (Mansir et al. 2018).

𝑀𝑊 =56,1×1000×3

𝑆𝑁 (3.1)

The FFA content was calculated from the acid value (𝐴𝑉, mg KOH/g) using Equation 3.2

(Demirbas 2009).

FFA =𝐴𝑉

2 (3.2)


59


Usually the solid catalysts are characterized by different instrumental techniques in order to

measure their morphology, physical properties and bulk properties. The catalytic behavior

depends on the morphological characteristics of the solid material, because the catalytic

process takes place at its surface (outer and inner). The most utilized techniques to

characterize materials’ morphology are BET and SEM. In terms of physical properties, the

surface area is the place of catalytic activity, but only a part is utilized in the catalytic reaction

(active center). In basic and acid catalysts, the active sites not only occupy a little fraction of

the surface, but also differ in basic and acid strength and sometimes in nature. Hammett

indicators are often used to determine the acid and basic strengths of a material. FTIR is useful

to identify the main chemical functional groups present on the surface of solid materials. For

bulk properties X-ray diffraction (XRD) is used to find: (i) the crystalline phases, (ii) crystalline

degree and (iii) crystallite size (Leofanti et al. 1997), with which the catalytic activity can be

interpreted.

In this work, the catalyst was prepared by drying the fly ash (FAD) for 2 h at 120 oC. Then, the

resulting material was characterized in terms of: (i) crystallographic structures, by powder X-

ray (XRD, PAN analytical Empyrean X-ray diffractometer equipped with Cu-Kα radiation source

λ= 1.54178 Å at 45 kV/ 40 mA); (ii) surface area (SBET) was estimated by the BET (Brunauer-

Emmett-Teller) method, pore size and pore volume were determined by the BJH (Barrett-

Joyner-Halenda) model. The specific surface area and pore structure characterization were

determined by nitrogen adsorption at 77 K using a surface area analyzer Micromeritics Gemini

V-2380. The samples were degassed overnight at 373 K before measurement; (iii) surface

morphology and quantitative analysis of elemental composition, by surface scanning electron

microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX), using a HR-FESEM Hitashi

SU-70 operated at 15 kV, equipped with a Bruker Quantax 400 EDS system; (iv) surface

functional species by Fourier transform infrared (FTIR, Agilent CARY 630 with wave number

range from 400 to 4000 cm−1); (v) basic and acid strength by using Hammett indicators (for

basic strength: neutral red, pKa = 6.8; bromothymol blue, pKa = 7.2; phenolphthalein,

pKa = 9.3; indigo carmine, pKa = 12.2; and 2,4-dinitroaniline, pKa = 15.0; indicators for acid


60

strength: bromothymol blue, pKa = 7.2; neutral red, pKa = 6.8; bromocresol purple, pKa = 6.1;

bromocresol green, pKa = 4.7; and bromophenol blue, pKa = 3.8). The latter method was

carried out by dispersing about 25 mg of the sample (catalyst) in 5.0 mL of a solution of

Hammett indicators (0.5 mg of indicator in 10 mL of methanol for basic strength or 10 mL of

benzene for acid strength), and left for 2 h to reach the chemical equilibrium. Then, the color

of the resulting solution was identified.

3.2.4 FAME synthesis and quantification

The experiments for FAME production were carried out in a batch reactor (in glass, 0.25 L of

capacity, equipped with temperature control and magnetic agitator), using 2 h of reaction time

at 600 rpm stirring speed.

At the end of each batch assay, the catalyst and methanol were separated from the reaction

mixture by centrifugation and evaporation, respectively. Then, the supernatant was placed

into a separating funnel over 12 h for phase separation. The water contained in the upper layer

of the liquid mixture was removed with anhydrous sodium sulfate (10 wt%) and weighed. The

resulting mixture, hereafter is so-called purified final mixture, was analyzed by gas

chromatography for FAME determination and was titrated with a KOH solution for final acid

value quantification (ICONTEC 218 1999).

The Shimadzu G-C 2014 chromatograph used for FAME determination was equipped with a

flame ionization detector and a capillary column SGEBP-20 60 m x 0.25 mm i.d. x 0.25 µm film

thickness with a stationary phase of polyethylene glycol; the carrier gas was helium with a flow

rate of 16.7 mL/min and a pressure of 2.5 atm; the injector (AOC-20i) was operated at 200 °C

and an injection volume of 2.0 µL in Split mode. Methyl heptadecanoate was used as internal

standard and hexane the solvent. The content of methyl esters was calculated based on the

standard method (UNE-EN ISO 14103) (AENOR-EN 14103 2011) and expressed as

concentration of FAME using the Equation 3.3:


𝐴𝐸𝐼×

𝑊𝐸𝐼

𝑊 (3.3)


61

Where 𝐶 is the concentration of FAME in the purified final mixture (w/w), ∑ 𝐴 is the total peak


heptadecanoate, 𝑊𝐸𝐼 is the mass (mg) of methyl heptadecanoate used and 𝑊 is the mass (mg)

of the sample used in the analysis.

The catalyst performance was assessed by the FAME yield and FFA conversion, calculated by

Equation 3.4 and Equation 3.5, respectively (Uprety et al. 2016; Wan Omar and Amin 2011a).

FAME yield (%) =𝐶 ×Total mass of purified final mixture

Mass of oil used in the experiment × 100 (3.4)

FFA conversion (%) = (1 −𝐴𝑉𝑓

𝐴𝑉𝑖) × 100 (3.5)

Where 𝐴𝑉𝑖 and 𝐴𝑉𝑓 correspond to the acid value of the initial oil mixture and of the purified

final mixture, respectively.

3.2.5 Experimental design and optimization of FAME production process

Response Surface Methodology (RSM) based on a Box–Behnken experimental design are a set

of mathematical and statistical techniques employed for designing experiments, creating

correlations, evaluating the effects of several factors, and their interaction effects for desirable

responses. This method uses the minimum required data that give the best reaction condition

for a desired response (Liu et al. 2014; Salamatinia et al. 2010) and was applied to optimize

and to investigate the relationship between operating conditions and the FAME yield. The

effect of four independent variables - catalyst loading, methanol/oil, RPO/WCO and reaction

temperature on the FAME yield was studied. The experimental range for each independent

variable (aka factor) tested in this work is shown in Table 3.1.


62

Table 3.1 - Range and factor levels of operating variables used in the Box – Behnken

experimental design.

Real variables Coded variables Level

Low (-1) Medium (0) High (+1)

Catalyst loading (wt%) A 5 10 15

Methanol/oil (molar ratio) B 3 6 9

Temperature (oC) C 45 50 55

RPO/WCO (wt%) D 0 (M3) 50 (M2) 100 (M1)

Twenty nine experimental runs were required, including five replicates of the central point.

The correlation in the form of a quadratic polynomial equation was developed for predicting

the response (i.e., FAME yield) as a function of independent variables and their interactions

according to Equation 3.6.

𝑌 = 𝛽0 + ∑ 𝛽𝑖𝑥𝑖𝑘𝑖=1 + ∑ 𝛽𝑖𝑖𝑥𝑖

2𝑘𝑖=1 + ∑ ∑ 𝛽𝑖𝑗𝑥𝑖𝑥𝑗

𝑘𝑗=𝑖+1

𝑘𝑖=1 + 𝜀 (3.6)

Where 𝑌 is the predicted response for the process, i.e., the dependent variable; 𝛽0 is the

intercept coefficient (offset); 𝛽𝑖 are the linear terms; 𝛽𝑖𝑖 are the quadratic terms; 𝛽𝑖𝑗 are the

interaction terms; 𝑥𝑖 and 𝑥𝑗 are the independent variables; and 𝜀 is the error (Hajamini et al.

2016).

Simplified regression models of Equation 3.6 (e.g., without interaction terms) were also fitted

to the experimental results. The best fit achieved with the simplest model was the one selected

and presented in this work.

The inference on the regression model was performed through an analysis of variance

(ANOVA), for a 95 % confidence level, where the statistically significant factors in the response

variable were identified, and an analysis of the coefficients of determination of the model, R2

and adjusted R2 (“Adj R2), was used to evaluate the adequacy of the regression model to the

experimental data. In this step one used the sum of the squares of residuals, instead of the

pure error.


63

Validation of the regression model assumptions (i.e., the assessment of the adequacy of the

model) was performed through a residual analysis (normality and residual plots). This analysis

was based on normalized/studentized residuals.

Once the best regression model was selected and validated, the optimal operating conditions

were identified through the response surface. Then, for statistical validity purposes, three runs

were performed using those optimal conditions, thus allowing to determine the deviations of

the data predicted by the model and the real values obtained experimentally. The software

Design – Expert 7.0.0 was used for the statistical data processing and analysis.

3.2.6 Catalyst reusability

Recovery, stability and reuse are important aspects of a heterogeneous catalyst to be applied

in biodiesel production. The reusability of FAD catalyst in esterification and transesterification

reactions was investigated through 2 successive catalytic cycles (i.e., in total 3 cycles) using the

optimal reaction conditions found in the optimization step. After each cycle, the solid catalyst

was recovered and activated by simple centrifugation, washing with isopropyl alcohol for

removing organic compounds eventually retained in the solid surface, calcined at 700 oC for

3 h and reused in the next catalytic cycle. At the end of each cycle, the catalyst was

characterized by XRD, textural properties and Hammett indicator.


64



The properties of the oil mixtures prepared for this study are shown in Table 3.2.

Table 3.2 - Properties of the oil mixtures used in this work.

M1 M2 M3

%WCO 0 50 100

%RPO 100 50 0

Moisture (wt%) 0.067±0.010 0.170 ± 0.003 0.197 ± 0.012

Density (g/mL) 0.908 ± 0.008 0.913 ± 0.010 0.906 ± 0.003

AV (mg KOH/g) 0.307 ± 0.004 3.958 ± 0.082 4.934 ± 0.252

FFA (wt%) 0.172 ± 0.005 1.979 ± 0.041 2.453 ± 0.056

MW (g/mol) 843.15 ± 9.52 886.34 ± 1.21 857.82 ± 4.01

Viscosity (mm2/s) @ 60 °C 14.902 ± 0.193 17.122 ± 0.123 19.185 ± 0.392

The properties of M1 (i.e., 100 % RPO) are similar to those reported by Kansedo et al. (2009)

and by Metawea et al. (2018). Concerning the waste cooking oils properties, they are quite

dependent of the vegetable oil feedstocks and their frying practices and conditions. The WCO

(M3) used in this work has properties similar to those reported by Wan Omar et al. (2011b)

and Lam and Keak (2010), and it can be categorized as yellow grease (FFA < 15 %) (Avhad and

Marchetti 2015).

As the percentage of WCO increases in the blend (see Table 3.2) higher are the moisture and

the FFA contents, the acid value and the viscosity, while the remaining properties values

(density and molecular weight) are similar among the three blends.


The solid catalyst prepared was characterized for some textural properties such as specific

surface area, crystalline structure, surface functional groups, but also their basic and acid

strength, etc. The results are shown and discussed in the next sections.


65

3.3.2.1 BET surface area and Hammett indicators analyses

FAD used in this work has an intermediate basic strength (10.1≤ pKb <12.2), due to the high

basicity of the metal-oxygen groups (Lewis bases) present in the calcium compounds on its

surface (see Sections 3.3.2.2 and 3.3.2.3) and a low acid strength (6.8≤ pKa<7.2) (Maneerung

et al. 2015). With regard to its textural properties, FAD has a low (BET) surface area (9.028

m2/g), characteristic of this type of material, a pore volume of 0.01055 cm3/g, and an average

pore diameter (77.188 Å), which shows some potential to the adsorption and desorption of

molecules such as triglycerides, glycerin and FAME (Chakraborty et al. 2010; Jacobson et al.

2008).

The average pore size distribution could be estimated from the nitrogen adsorption-

desorption isotherms. Figure 3.1 shows those isotherms for FAD catalyst, which behaves like a

type IV(a) according to the classification of the International Union of Pure and Applied

Chemistry. The initial part of this graph exhibits a behavior such as the type II isotherm, typical

of monolayer adsorption. Subsequently, a hysteresis cycle associated with the characteristic

capillary condensation of mesoporous solids is observed, which is observed for pore size

ranges between 20-500 Å (Thommes et al. 2015).

Figure 3.1 - Adsorption and desorption isotherms for the FAD catalyst.


66

3.3.2.2 SEM and EDX analyses

The SEM images for characterizing the morphological characteristics and EDX for elemental

analysis and chemical characterization of the catalysts were obtained. Figure 3.2 shows the

morphological and the elemental composition of FAD catalyst. Ash particles have uniform

distribution of agglomerates with irregular shapes and rough structure. The same

characteristics were observed by Rajamma et al. (2009). In addition, the average particle size

for the FAD catalyst was 4.353 μm ( 1.07) as determined using the ImageJ software. The

results of EDX show as predominant elements: Ca, Mg, Si, Al, O, K, S, Na, Cl and P

(Figure 3.2 c).

(a) (b)

(c)

Figure 3.2 - FAD catalyst characterization by: SEM (a and b) and EDX (c).


67

3.3.2.3 XRD analyses

The XRD diffractogram of the FAD catalyst is depicted in Figure 3.3. The XRD pattern shows

clear diffraction peaks corresponding to calcium oxide (CaO) phase detected at 2=32.2o,

37.4o, 53.8o, 65.2o, and 67.5o, calcium carbonate (CaCO3-major component) phase detected at

2 =23.3o, 29.6o, 36.2o, 39.7o, 43.4o, 47.8o, 48.8o, 56.9o, 61.0o and 65.0o, potassium chloride

(KCl) phase detected at 2 =28.5o, 40.5o, and silicon dioxide (SiO2) phase detected at 2 =20.9o,

26.7o, 36.38o, 39.46o, 40.28o, 50.2o, 60.2o and 68.5o, among other components such as calcium

hydroxide detected at 2=17.91°, 28.51°, 33.95°, 47.41°, 50.68°, 64.60°.

Figure 3.3 - XRD patterns of FAD catalyst.

Regarding the semi-quantitative mass composition, a high content of calcium carbonate

(71.0 %) was found, followed by calcium hydroxide (12.9 %), potassium chloride (7.1 %),

calcium oxide (3.8 %), silicon dioxide (2.3 %) and other components in smaller proportion were

identified (3.0 %); similar compounds were reported by Sharma et al. (2012) for wood ash and

by Ho et al. (2014) for palm oil mill fly ash. The presence of calcium hydroxide may be due to

the ambient humidity that reacts (after the combustion process) with calcium oxides presents

on the surface of the solid; this phenomenon was also observed by Maneerung et al. (2015).

The high calcium carbonate content results from the carbonation of calcium oxides and

hydroxides since solid material was in contact with atmospheric carbon dioxide, after the

combustion process.

0

2000

4000

6000

8000

10000

12000

15 20 25 30 35 40 45 50 55 60 65

Intensity(u.a)

2ϴ

CaCO3

Ca(OH)2

KCl

CaO

SiO2


68

3.3.2.4 FTIR analyses

The FTIR spectrum of FAD used in this work is shown in Figure 3.4. It shows the major

absorption broad band at 1408.1 cm-1 and minor absorption bands at 872 and 712 cm-1, which

correspond to the asymmetric stretching and to out-of-plane band and in-plane band vibration

modes of carbonate (CO3-2) group, respectively. The small bands at 2510 and 2320 cm-1 also

correspond to the characteristic spectrum of this functional group. This result confirms the

presence of CaCO3 in FAD, detected by XRD.

Figure 3.4 - FTIR spectrum of FAD catalyst.

PO4-3 and Si-O components (silica phosphates) show broad bands in the region between 1138

and 942 cm-1; the same was observed by Maneerung et al. (2015) and Sharma et al. (2012) in

bottom ash waste from woody biomass gasification and wood ash from the combustion of

Acacia nilotica (babul), respectively. Moreover, the absorption sharp band at 3642 cm-1, which

is attributed to -OH band, was observed in the FAD catalyst, this band is in agreement with the

presence of Ca(OH)2 as determined by XRD, and an evidence of the possible water absorption

on the CaO surface producing Ca(OH)2 (Boey et al. 2011).

3.3.3 Optimization of FAME production process: regression model and statistical analysis

The experimental results obtained in the set of assays aiming at optimizing the FAME yield are

shown in Table 3.3.


69

Table 3.3 - Experimental design and predicted results of RSM.

Run

Real variables FAME yield (wt%)

Catalyst loading Methanol/oil T RPO/WCO Experimental Predicted*

(wt%) (mol/mol) (°C) (wt%)

1 15 6 50 100 11.40 21.35

2 10 6 55 0 60.94 63.48

3 10 3 50 0 41.44 41.84

4 10 6 50 50 62.86 55.95

5 10 9 55 50 64.64 62.18

6 15 6 45 50 25.82 31.83

7 10 3 45 50 12.25 17.16

8 10 3 50 100 0.00 6.68

9 5 9 50 50 38.33 35.43

10 10 6 45 0 25.24 26.45

11 5 6 50 100 8.96 1.57

12 15 6 55 50 65.35 68.86

13 10 6 50 50 70.34 55.95

14 5 6 55 50 47.94 49.07

15 10 6 55 100 27.57 28.32

16 10 6 45 100 7.39 -8.70

17 15 3 50 50 52.83 47.22

18 10 9 45 50 33.25 25.16

19 10 9 50 100 8.57 14.68

20 5 3 50 50 28.34 27.43

21 5 6 45 50 0.00 12.05

22 10 9 50 0 40.32 49.83

23 15 6 50 0 68.20 56.51

24 10 6 50 50 38.63 55.95

25 10 3 55 50 59.65 54.18

26 10 6 50 50 45.84 55.95

27 15 9 50 50 57.38 55.21

28 5 6 50 0 38.70 36.73

29 10 6 50 50 62.08 55.95

*Predicted by Equation 3.7


70

The regression model of Equation 3.6 fitted to the experimental results revealed that the

interaction between the factors 𝛽𝑖𝑗𝑥𝑖𝑥𝑗 was not significant (p-value > 0.05). Thus, the

simplified model (i.e., quadratic model without interactions) was used and the goodness-of-fit

was evaluated by the several parameters determined in the ANOVA. The results are shown in

Table 3.4, where it can be seen that model has a good fit as R2=0.8702 and Adj R2=0.8182. The

R2 value indicates that 87.02 % of the variability in the data is predicted by the model.

Table 3.4 - ANOVA results of the response surface quadratic model without interactions.

Source of variations

Sum of squares

Degrees of freedom

Mean square

F - value p – value

Model 12028.36 8 1503.55 16.75 < 0.0001

Residual 1794.79 20 89.74

Lack of fit 1100.2 16 68.76 0.936 0.9183

Pure error 694.59 4 173.65

Total 13823.15 28

R2 = 0.8702 Adj Pred C.V.a = 24.88% S.D.b = 9.47

R2 = 0.8182 R2 = 0.7424

a C.V.= coefficient of variation.

d S.D.= standard deviation.

The low p-value (< 0.0001) of the model means that it is statistically significant. On the other

hand, the lack of fit F-value of 0.936 implies that is not significant relative to the pure error,

i.e. the lack of fit of the model is statistically non-significant; there is a 91.83 % chance that this

value could occur due to noise. The “Pred R2” of 0.7424 is in reasonable agreement with the

“Adj R2” value of 0.8182. In short, the selected regression model satisfactorily predicts the

effect of the four factors on FAME yield. Equation 3.7 represents the model developed:

𝑌 = 55.95 + 9.89 𝑥𝐴 + 4.00 𝑥𝐵 + 18.51 𝑥𝐶 − 17.58 𝑥𝐷 − 6.92 𝑥𝐴2 − 7.70 𝑥𝐵

2 − 8.57𝑥𝐶2 − 19.98𝑥𝐷

2 (3.7)

Where 𝑌 is the response variable (FAME yield, wt%), 𝑥𝐴 (catalyst loading, wt%), 𝑥𝐵

(methanol/oil, molar ratio), 𝑥𝐶 (reaction temperature, oC) and 𝑥𝐷 (RPO/WCO, wt%) are the

studied factors. The positive sign of a coefficient term means synergistic effect while the


71

negative sign reveals the opposite effect, of the influencing variables on FAME yield yield (Liu

et al. 2014). The FAME yield predicted by this regression model is shown in Table 3.3, for

comparison with experimental results.

The statistical significance of each regression coefficient of the model on the response variable

was evaluated using ANOVA testing and the results are shown in Table 3.5. The p-values

indicate the significance of each regression coefficient. In general, smaller p-value (< 0.05)

indicates higher significance of the corresponding coefficient (Avramović et al. 2010).

According to obtained results, three of the four linear factors were statistically significant

(𝑥𝐴 , 𝑥𝐶 , and 𝑥𝐷) and only one (𝑥𝐵) was not significant (for confidence level of 95 %). Besides

that, the influence of square value of RPO/WCO ( 𝑥𝐷2 ) with a negative effect of -19.98 (p-value

< 0.0001) was found to be the most significant term affecting the FAME yield; the quadratic

term of the temperature was also significant (p-value = 0.0320).

Table 3.5 - ANOVA results for the coefficients of the variables in the quadratic regression model

without interactions.

Model parameters Estimate coefficient F - value p - value

Intercept 55.95

𝑥𝐴 9.89 13.09 0.0017

𝑥𝐵 4.00 2.14 0.1592

𝑥𝐶 18.51 45.82 < 0.0001

𝑥𝐷 -17.58 0.158 < 0.0001

𝑥𝐴2 -6.92 3.46 0.0775

𝑥𝐵2 -7.70 4.29 0.0515

𝑥𝐶2 -8.57 5.31 0.0320

𝑥𝐷2 -19.98 28.88 < 0.0001

In order to validate the assumptions of the simplified regression model (i.e., quadratic model

without interactions), statistical graphical methods were used. A normal probability plot of

residuals is shown in Figure 3.5 a, which corresponds to the difference between the


72

experimental and the predicted response. The data points are located approximately along a

straight line, thus one can intuitively conclude that the residuals follow a normal distribution.

Plot of residuals versus fitted response values (predicted) is depicted in Figure 3.5 b, which

shows that the residuals are randomly distributed. Residuals are located in a horizontal line

and the number of points that exist in the above and below of horizontal line is equal.

Moreover, residual values are in the range ± 3.00; typically, a threshold of three standard

deviations is employed as a definition of an outlier (Noshadi et al. 2012). The actual FAME yield

versus the predicted values is plotted in Figure 3.5 c, which corroborates the goodness-of-fit

of the regression model developed. In brief, this analysis confirms the accuracy and reliability

of the proposed regression model.


73

(a) (b)

(c)


(c) Predicted versus actual values plot.

As stated before, in the range tested (Table 3.1), the factors studied in this work, except

methanol/oil molar ratio, had a statistically significant influence on FAME yield; although some

were more significant than others. This is shown as response surface plots in Figure 3.6.

Externally Studentized Residuals

No

rmal %

Pro

bab

ilit

y

Normal Plot of Residuals

-3,00 -2,00 -1,00 0,00 1,00 2,00 3,00

1

5

10

20

30

50

70

80

90

95

99

Yield

Color points by value of

Yield :

0,00 70,34

Predicted

Exte

rnally S

tud

en

tize

d R

esi

du

als

Residuals vs. Predicted

-4,00

-2,00

0,00

2,00

4,00

-20,00 0,00 20,00 40,00 60,00 80,00

3,64462

-3,64462

0

Yield


Yield :

0,00 70,34

Actual

Pre

dic

ted

Predicted vs. Actual

-20,00

0,00

20,00

40,00

60,00

80,00

-20,00 0,00 20,00 40,00 60,00 80,00

Yield


Yield :

0,00 70,34


74

(a) (b)

(c) (d)

(e) (f)

Figure 3.6 - Response surface plots of FAME yield as a function of: (a) RPO/WCO ratio and catalyst loading

at 50 oC and methanol/oil = 6 mol/mol; (b) RPO/WCO ratio and methanol/oil at 50 oC and catalyst loading

= 10 wt%; (c) RPO/WCO ratio and temperature for catalyst loading = 10 wt% and methanol/oil = 6

mol/mol; (d) temperature and methanol/oil ratio for catalyst loading = 10 wt% and RPO/WCO = 50 wt%;

(e) temperature (oC) and catalyst loading (wt%) for methanol/oil = 6 mol/mol and RPO/WCO = 50 wt%; (f)

methanol/oil ratio and catalyst loading (wt%) at 50 oC and RPO/WCO = 50 wt%.

0

20

40

60

80

100

56

78

910

1112

1314

15

020

4060

80

FA

ME

Yie

ld (

wt%

)

Cat

alys

t loa

ding

(wt%

)

RPO/WCO (wt%)

0

<20 %

20-40 %

40-60 %

60-80 %

80-100 %

0

20

40

60

80

100

3

4

5

6

7

8

9

020

4060

80

FA

ME

Yie

ld (

wt%

)

Met

hano

l/oil (m

ol/m

ol)

RPO/WCO (wt%)

0

<20 %

20-40 %

40-60 %

60-80 %

80-100 %

0

20

40

60

80

100

44

46

48

50

52

54

56

020

4060

80

FA

ME

Yie

ld (

wt%

)

Tem

pera

ture

(ºC

)

RPO/WCO (wt%)

0

<20 %

20-40 %

40-60 %

60-80 %

80-100 %

0

20

40

60

80

100

3

4

5

6

7

8

9

4446

4850

5254

FA

ME

Yie

ld (

wt%

)

Met

hano

l/oil (m

ol/m

ol)

Temperature (ºC)

0

<20 %

20-40 %

40-60 %

60-80 %

80-100 %

0

20

40

60

80

100

56

78

910

1112

1314

15

4446

4850

5254

FA

ME

Yie

ld (

wt%

)

Cat

alys

t loa

ding

(wt%

)

Temperature (ºC)

0

< 20 %

20-40 %

40-60 %

60-80 %

80-100 %

0

20

40

60

80

100

56

78

910

1112

1314

15

34

56

78

FA

ME

Yie

ld (

wt%

)

Cat

alys

t loa

ding

(wt%

)

Methanol/oil (mol/mol)

0

<20 %

20-40 %

40-60 %

60-80 %

80-100 %


75

From Figure 3.6 a, higher yields of FAME (64 %) were achieved with high catalyst loading

(13.2 wt%) and moderate (28.0 wt%) RPO/WCO mass ratio. For any fixed RPO/WCO mass ratio,

as catalyst loading increased higher FAME yields were observed, which may be due to the

higher number of active sites (of the catalyst) available in the reaction medium. On the other

hand, RPO/WCO mass ratio higher that 28.0 wt% affected negatively the FAME yield for any

catalyst loading tested. Thus, the low acid strength and intermediate basic strength of FAD

seems to be suitable to catalyze oily mixtures with higher FFA contents, which, according to

some authors (Rabiah et al. 2014; Wan Omar et al. 2011b), could be due to the balance of acid

and basic catalyst sites. This is a promising result for the economic and environmental

sustainability of the process. Concerning methanol/oil molar ratio, Figure 3.6 b shows the weak

influence of this factor on the response variable; therefore, the methanol/oil molar ratio of 6.6

can be used to achieve the highest yields to FAME (64 %). Similar behavior was mentioned by

Volli et al. (2019) for the same methanol/oil molar ratio but using bone impregnated fly ash as

a catalyst. Figure 3.6 c shows the influence on FAME yield of the two most significant factors

studied in this work: temperature and RPO/WCO mass ratio. Indeed, increasing the reaction

temperature rose the FAME yield independently of the RPO/WCO mass ratio used. The higher

yield was observed at 55 oC with 28.0 wt% of RPO/WCO mass ratio (loading = 10 wt% and

methanol/oil = 6 mol/mol); Uprety et al. (2016) also found a very significant effect between 50

and 60 oC (reaction temperature) on the yield, using a catalyst of CaO and RPO as raw material.

Figure 3.6 d shows once more the different relevance of the temperature and methanol/oil

molar ratio on the response variable. Catalyst loading and reaction temperature had similar

positive effects on the FAME yield (Figure 3.6 e), being the higher yields achieved (c.a. 74 %)

at 55 oC and catalyst loading 13.2 wt% (for methanol/oil molar ratio of 6 and RPO/WCO mass

ratio of 50 wt%). This high FAME yield achieved may be due to the crystalline phases (calcium

hydroxide and calcium oxide), the functional groups (carbonate group) and pore diameter

(average 77.188 Å) found in the solid catalyst.

Regarding the percentage of conversion of free fatty acids, similar results were obtained for

the different experiments with values close to 84.3 % ± 6.0 %; which may be due to the slightly

acid character of the solid catalyst (6.8≤ pKa<7.2). These conversion values point out to a


76

bifunctional character of the FAD catalyst, i.e., simultaneous catalysis of transesterification and

esterification reactions, already found by Vargas et al. (2019).

Optimal operating condition

An important objective of this study was to find optimal operating conditions to achieve

maximum FAME yield, combining the several independent variables studied. The RMS

suggested that the highest FAME yield was 73.8 %, which can be achieved by using 13.57 wt%

of catalyst loading, 6.7 methanol/oil molar ratio, 28.04 wt% of RPO/WCO mass ratio and 55 oC

for the reaction temperature. To validate the proposed operating conditions, three replicate

experiments were conducted under them, over 2 h at 600 rpm stirring speed. The average

experimental FAME yield was 78.8 % ( 1.7 %), which is close to the predicted value

(i.e., 73.8 %). So, the validity of the proposed correlation is confirmed again with an error of

6.8 % ( 0.05 %). The FAME yield and the respective relative error between model predictions

and experimental value were close to the obtained by Badday et al. (2013), using activated

carbon-supported tungstophosphoric as catalyst on the Jatropha oil and a Central Composite

Design (CCD) as experimental design method.

3.3.4 Catalyst reusability: catalytic performance assessment

The reusability of a catalyst is very important for its commercial feasibility. In order to

investigate the reusability of the FAD catalyst, the subsequent reaction cycles were carried out

under the optimized reaction condition: 13.57 wt% of catalyst loading, 6.7 methanol/oil molar

ratio, 28.04 wt% of RPO in the oil mixture and 55 oC for the reaction temperature, 2 h reaction

time and 600 rpm stirring speed. Between each cycle the catalyst was regenerated, according

to the procedure stated in Section 3.2.6. The FAME yields obtained from the reused catalyst in

each cycle is shown in Figure 3.7, where a slight increase on FAME yield with the repeated

usage of the catalyst is observed. However, a statistical analysis of the data (ANOVA for a

confidence level of 95 %) showed that differences observed among the three assays were not

statistically significant with p-value = 0.1258; therefore, the activity of the FAD catalyst could

be considered roughly constant over three cycles of use. Similar catalytic stability were


77

reported by Chakraborty et al. (2010) and Maneerung et al. (2015) using fly ash from a thermal

power plant with a combustion technology and bottom ash waste arising from woody biomass

gasification, respectively.

Figure 3.7 - Reusability studies of the FAD catalyst under the optimal operating conditions.

The XRD patterns of the reused catalyst after each regeneration cycle are shown in Figure 3.8.

It can be observed in the superimposed diffractograms that the majority phase was calcium

carbonate (CaCO3), phase detected at 2=23.3o, 29.6o, 36.2o, 39.7o, 43.4o, 47.8o, 48.8o and

56.9o, followed by calcium hydroxide (Ca(OH)2) detected at 2=17.91°, 33.95° and 50.68°,

indicating that CaO was partially transformed into Ca(OH)2, probably through the reaction of

CaO (2=32.2°, 37.4° and 53.8°) with H2O in small amount in the reactants and/or moisture

during the repeated usage of catalyst (Maneerung et al. 2015). This may explain the observed

slight increase of FAME yield over the reuse cycles of the FAD, as Ca(OH)2 has catalytic

properties. A peak is also observed in 2= 26.7o due to the presence of the phase silicon dioxide

(SiO2) that gives the catalyst a low acid strength (and thus a bifunctional). The KCl (2=28.5o

and 40.5o) found did not contribute to the catalytic activity of the FAD. This conclusion arose

from three experimental tests performed with pure KCl as a catalyst and where no FAME yield

was registered (results not shown).

73.5%76.2%

83.6%

0%

20%

40%

60%

80%

100%

Fresh Cycle 1 Cycle 2

FAM

E yi

eld

(w

t%)


78

Figure 3.8 - XRD patterns of FAD catalyst for the different reuse cycles.

The BET surface area, pore volume, pore diameter and basic and acid strength of reused

catalyst (FAD) are shown in Table 3.6. The basic and acid strength of this reused catalyst and

its textural properties did not change throughout the reuse cycles and regeneration steps.

Table 3.6 - Textural properties and acid/basic strength of the FAD catalyst used in three FAME

synthesis cycles.

Catalyst sample

Specific surface

area (m2/g)

Pore volume (cm3/g)

Average pore

diameter (Å)

Basic strength Acid strength

Fresh 9.0280 0.01055 77.188 10.1≤ pKa <12.2 6.8≤ pKa<7.2

Cycle 1 10.9496 0.01253 82.639 10.1≤ pKa <12.2 6.8≤ pKa<7.2

Cycle 2 10.2876 0.01147 80.365 10.1≤ pKa <12.2 6.8≤ pKa<7.2

From these results, it is reasonable to conclude that after the cycles of reuse of the catalyst, it

did not lose its catalytic activity (FAME yield). Indeed, the catalytic activity seems to be slightly

increased, although not statistically different between the tests done, which can be explained

due to relatively small changes of its textural properties, crystalline active phases (CaCO3,

Ca(OH)2 and SiO2) and basic and acid strength (surface chemistry) throughout the reuse cycles.


79

Thus, FAD catalyst has shown good properties for reuse in the process, and that is an

advantage, because it is a low cost material, produced from an industrial waste and thus can

turn the process more sustainable in terms of natural resources integration.

3.4 Conclusions

An efficient fly ash residual catalyst (FAD) was evaluated with mixtures of RPO and WCO to

produce FAME using the response surface methodology and an experimental design Box

Behnken type for optimizing the response variable (FAME yield). A regression quadratic model

without interactions was the one that best fitted the experimental results, predicting the

following optimal operating conditions: catalyst loading of 13.57 wt%, methanol/oil molar

ratio of 6.7, RPO in the oil mixture of 28.04 wt%, temperature of 55 oC in batch regime over

2 h and 600 rpm of stirring speed. Under these operating conditions maximum FAME yield

expected is 78.8 %. In the tested ranges, the most significant variables (95 % confidence level)

affecting the FAME yield were the RPO/WCO mass ratio and the reaction temperature (oC),

both with p-value <0.0001, followed by the catalyst loading (p-value = 0.0017). On the other

hand, the methanol/oil molar ratio was not significant (p-value = 0.1592), indicating that the

lowest ratio tested can be used to achieve the higher FAME yield registered.

The selected regression model accurately predicted the experimental results with a R2 = 0.8702

and Adj R2 = 0.8182. Three assays were carried out under the optimal operating conditions,

where the average of FAME yield reached was 78.8 % ( 1.7 %), near the predicted by the

regression model (73.8 %). Thus, the validity of the proposed regression model was

demonstrated.

This works showed that FAD catalyst can be used for up to three cycles without loss of catalytic

activity. However, the catalyst should be regenerated between each cycle, by washing with

isopropyl alcohol and calcined at 700 oC for 3 h. The characterization of the surface, textural

and crystalline properties of the catalyst, after use in each FAME synthesis cycle, showed that

those properties were not significantly affected. The acid and basic strength remained


80

constants. In addition, one recommends to evaluate the reuse of FAD above three cycles in

order to find the maximum number of cycles that it could be used keeping a high FAME yield

and carry out a complementary characterization of the acid and basic catalyst sites with

temperature-programmed desorption of NH3 and CO2 techniques, respectively.

Summing up, exploiting residual feedstocks, this work gives a sustainable and affordable

approach to lower the biodiesel production costs and simultaneously, minimizing the

environmental burdens traditionally inherent to the management of two wastes streams:

WCO and FAD.

Therefore, an awareness should be created so that any material that is deemed a waste could

be exploited for usage in this or other applications, thereby implementing the principles of

circular economy.

Acknowledgments

Edgar M. Vargas S. expresses his sincere gratitude to the Universidad Jorge Tadeo Lozano

(Direction of Investigation, Creation and Extension) for the financial assistance of this work.

Márcia C. Neves acknowledges FCT, I.P. for the research contract CEECIND/00383/2017 under

the CEEC Individual 2017. The authors thank for the financial support to FCT/MCTES for the

financial support to CESAM (UIDP/50017/2020 & UIDB/50017/2020) and CICECO

(UIDB/50011/2020 & UIDP/50011/2020) through national funds.


81

References


ASTM. 2010. “Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity).” Annual Book of ASTM Standards: 1–10.


Avramović, Jelena M. et al. 2010. “The Optimization of the Ultrasound-Assisted Base-Catalyzed Sunflower Oil Methanolysis by a Full Factorial Design.” Fuel Processing Technology 91(11): 1551–57.

Badday, Ali Sabri, Ahmad Zuhairi Abdullah, and Keat Teong Lee. 2013. “Optimization of Biodiesel Production Process from Jatropha Oil Using Supported Heteropolyacid Catalyst and Assisted by Ultrasonic Energy.” Renewable Energy 50: 427–32. http://dx.doi.org/10.1016/j.renene.2012.07.013.

Boey, Peng Lim et al. 2011. “Utilization of BA (Boiler Ash) as Catalyst for Transesterification of Palm Olein.” Energy 36(10): 5791–96. http://dx.doi.org/10.1016/j.energy.2011.09.005.

Boey, Peng Lim, Gaanty Pragas Maniam, and Shafida Abd Hamid. 2009. “Biodiesel Production via Transesterification of Palm Olein Using Waste Mud Crab (Scylla Serrata) Shell as a Heterogeneous Catalyst.” Bioresource Technology 100(24): 6362–68. http://dx.doi.org/10.1016/j.biortech.2009.07.036.

Chakraborty, R., S. Bepari, and A. Banerjee. 2010. “Transesterification of Soybean Oil Catalyzed by Fly Ash and Egg Shell Derived Solid Catalysts.” Chemical Engineering Journal 165(3): 798–805. http://dx.doi.org/10.1016/j.cej.2010.10.019.

Chatterjee, Amrita, Xijun Hu, and Frank Leung Yuk Lam. 2018. “Catalytic Activity of an Economically Sustainable Fly-Ash-Metal-Organic- Framework Composite towards Biomass Valorization.” Catalysis Today 314(October 2017): 137–46. https://doi.org/10.1016/j.cattod.2018.01.018.

Chen, Jein Wen et al. 2012. “Carcinogenic Potencies of Polycyclic Aromatic Hydrocarbons for Back-Door Neighbors of Restaurants with Cooking Emissions.” Science of the Total Environment 417–418: 68–75. http://dx.doi.org/10.1016/j.scitotenv.2011.12.012.



82

Girón, R. P. et al. 2015. “Adsorbents/Catalysts from Forest Biomass Fly Ash. Influence of Alkaline Activating Agent.” Microporous and Mesoporous Materials 209: 45–53. http://dx.doi.org/10.1016/j.micromeso.2015.01.051.







Kotwal, M. S. et al. 2009. “Transesterification of Sunflower Oil Catalyzed by Flyash-Based Solid Catalysts.” Fuel 88(9): 1773–78. http://dx.doi.org/10.1016/j.fuel.2009.04.004.


Leofanti, G et al. 1997. “Catalyst Characterization: Characterization Techniques.” Catalysis Today 34: 307–27. https://ac-els-cdn-com.ezproxy.unal.edu.co/S0920586196000569/1-s2.0-S0920586196000569-main.pdf?_tid=19d5e4b8-7ff0-4112-8d38-dcc4bfd030cd&acdnat=1550332601_264d3d96907b4cb6bcc5cdfbeff93007 (February 16, 2019).



83




Mendonça, Iasmin M. et al. 2019. “New Heterogeneous Catalyst for Biodiesel Production from Waste Tucumã Peels (Astrocaryum Aculeatum Meyer): Parameters Optimization Study.” Renewable Energy 130: 103–10.

Metawea, Rodaina et al. 2018. “Process Intensification of the Transesterification of Palm Oil to Biodiesel in a Batch Agitated Vessel Provided with Mesh Screen Extended Baffles.” Energy 158: 111–20.

Noshadi, I., N. A.S. Amin, and Richard S. Parnas. 2012. “Continuous Production of Biodiesel from Waste Cooking Oil in a Reactive Distillation Column Catalyzed by Solid Heteropolyacid: Optimization Using Response Surface Methodology (RSM).” Fuel 94: 156–64. http://dx.doi.org/10.1016/j.fuel.2011.10.018.


Rabiah Nizah, M. F. et al. 2014. “Production of Biodiesel from Non-Edible Jatropha Curcas Oil via Transesterification Using Bi2O3-La2O3 Catalyst.” Energy Conversion and Management 88: 1257–62. http://dx.doi.org/10.1016/j.enconman.2014.02.072.

Rajamma, Rejini et al. 2009. “Characterisation and Use of Biomass Fly Ash in Cement-Based Materials.” Journal of Hazardous Materials 172(2–3): 1049–60.




84

Thommes, Matthias et al. 2015. “IUPAC Technical Report Physisorption of Gases, with Special Reference to the Evaluation of Surface Area and Pore Size Distribution (IUPAC Technical Report).” https://www.3p-instruments.com/wp-content/uploads/2017/04/2015-IUPAC-Technical-Report.pdf (February 18, 2019).


Vargas, Edgar M., Márcia C. Neves, Luís A.C. Tarelho, and Maria I. Nunes. 2019. “Solid Catalysts Obtained from Wastes for FAME Production Using Mixtures of Refined Palm Oil and Waste Cooking Oils.” Renewable Energy 136: 873–83.

Volli, Vikranth, Mihir Kumar Purkait, and Chi Min Shu. 2019. “Preparation and Characterization of Animal Bone Powder Impregnated Fly Ash Catalyst for Transesterification.” Science of the Total Environment 669: 314–21.


Wan Omar, Wan Nor Nadyaini, and Nor Aishah Saidina Amin. 2011b. “Optimization of Heterogeneous Biodiesel Production from Waste Cooking Palm Oil via Response Surface Methodology.” Biomass and Bioenergy 35(3): 1329–38. http://dx.doi.org/10.1016/j.biombioe.2010.12.049.

85

SECTION D - Optimization of FAME production in continuous fixed-bed reactor

In section D the design and construction of a continuous fixed bed reactor (CFBR) is presented

and the study to optimize the FAME production process with methanol at 60 oC, using mixtures

of WCO and RPO and pelletized biomass fly ashes as chosen solid catalyst (section B). The

effect of three operating variables (residence time, WCO/RPO mass ratio and methanol/oil

molar ratio) on FAME concentration was studied, using the experimental Box-Benhken design

and the Response Surface Methodology. Some preliminary assays were carried out to identify

the time at which the steady state of the continuous fixed bed reactor used in this work was

reached. Seventeen experimental runs were required, including five replicates of the central

point); all were carried out using a fixed bed depth of 15 cm (bulk density: 2.16 g/mL) and an

operating time of 9 h. Furthermore, the solid catalyst was characterized by SEM, EPS, XRD, BET,

FT-IR and Hammett indicators. Additionally, the pelletized catalytic stability of the biomass fly

ash fixed bed was evaluated through an assay during 32 h of operation, where the FAME

concentration was monitored over time. The optimal operating conditions found by the

regression model were used in this assay.

The information presented in this section was adapted from the following published

article:

E. M. Vargas, Duvan O. Villamizar, M. C. Neves, and M. I. Nunes, “Pelletized biomass fly ash

for FAME production: optimization of a continuous process”, Fuel, vol. 293, pp. 120425,

2021, doi: https://doi.org/10.1016/j.fuel.2021.120425.


86

OPTIMIZATION OF FAME PRODUCTION IN A CONTINUOUS FIXED-BED REACTOR

87

4 Pelletized biomass fly ash for FAME production: optimization of a

continuous process

Abstract: Circularity in the resources usage is one of the current challenges in the development

of our civilization. At the same time, there is an imperative need for cleaner and competitive

energy sources, alternative to those of fossil origin. In this context, the present work aimed to

optimize a continuous process for fatty acid methyl esters (FAME) production using residual

resources, namely waste cooking oil (WCO) and pelletized biomass fly ash as catalyst. A

continuous fixed bed reactor was designed and built. The pelletized catalyst performance was

assessed in the reaction system, which was fed with a mixture of refined palm oil (RPO), WCO

and methanol at 60 °C. The effect of three operating variables (residence time, WCO/RPO mass

ratio and methanol/oil molar ratio) on FAME concentration was studied, using experimental

the Box-Benhken design and the Response Surface Methodology (RSM). The maximum FAME

concentration achieved was c.a. 89.7 % under the following operating conditions: 124 min of

residence time, 74.6 wt% WCO/RPO and 12:1 methanol/oil molar ratio. The catalyst kept

stable over the 32 h of continuous operation, without noticeable deactivation. Thus, the

pelletization of an industrial biomass fly ash, with bifunctional catalytic properties, allowed its

application to the FAME production in a continuous regime, with high performance even when

high percentages of WCO were used as feedstock.

Keywords: Biomass fly ash; continuous fixed bed reactor; heterogeneous catalysts; FAME;

response surface methodology; waste cooking oil.


88

4.1 Introduction

In order to meet the current and growing demand for liquid fuels as well as the increasing

concerning about climate change, biodiesel has been pointed out as a sustainable solution to

partially replace fossil diesel (Suarez et al. 2009). Biodiesel is synthesized through the

transesterification and esterification reactions of vegetable oils, animal fats or waste vegetable

oil using alcohols and basic and acid catalysts (homogeneous, heterogeneous and enzymatic)

(Fang et al. 2011; Mittelbach et al. 1983; Pirez et al. 2012).

Some residual biomass fly ashes have been proven to have catalytic properties suitable for

biodiesel production (Vargas et al. 2019a). Uprety et al. (2016) evaluated wood ash from two

different sources (birch bark and fly ash from a biomass-based power plant) in the biodiesel

production; 88.06 % was the highest yield to FAME observed with birch bark ash. Sharma et

al. (2012) studied calcined wood ash and activated wood ash as catalysts for the production of

biodiesel using Jatropha oil obtaining a conversion in the range of 97-99 %. In both studies the

wood ashes contained a high dispersion of CaCO3, CaO, Ca(OH)2 and SiO2 (active phase)

pointed out as responsible for the observed catalytic performance.

There are several research works (e.g. Vargas et al. (2019a), Teixeira et al. (2019), Miller et al.

(2006) and Vargas et al. (2019b)) that have been dedicated to the valorization of residual

materials for biodiesel production, aiming to reduce the production costs but also to

operationalize the principles of circular economy. Those works (e.g. Tarelho et al. (2015),

Jensen et al. (2004)) focused both in the preparation of residual materials with catalytic

properties and in the use of residual vegetable oils and fats.

Currently, the biodiesel production at industrial scale comprises three main stages: (i) reaction

stage carried out in batch, semi-continuous or continuous reactors, (ii) settling stage for light

phase and glycerin separation and (iii) purification stage of the light phase, to obtain a purified

biofuel. The reaction stage is carried out at moderate conditions between 50.7 – 405.3 kPa and

30 – 90 °C, using a homogeneous basic catalyst (Evangelista et al. 2012) or acid catalyst when

the vegetable oils have high free fatty acids (FFA) content. However, the basic catalysts can


89

form soaps in the presence of FFA and water, negatively affecting the efficiency of the

separation stage (Freedman et al. 1981; Zhang et al. 2003). Other hotspot of the conventional

biodiesel production process is the high energy consumption in the separation stage (if

centrifuges are used), besides the contamination of large quantities of water during the

purification stage contributing for the high water footprint of the process (Vyas et al. 2010).

Nowadays, in some industrial processes, the continuous flow and fixed bed reactors are the

most widely used due to their high production capacity and easy process control (Andrigo et

al. 1999; Tran et al. 2017). Therefore, this type of reactors can be considered as a good

alternative to produce biodiesel at industrial scale. At lab scale, Park et al. (2008) used a

continuous fixed bed reactor packed with a solid catalyst of WO3/ZrO for oleic acid

transesterification; the conditions used were 75 °C, methanol/oil ratio of 20:1 and an operation

time of 140 h; the FAME yield achieved was 85 %. Moreover, Da Silva et al. (2014) evaluated

the behavior of zinc oxide and aluminum oxide as catalysts in biodiesel production, using

soybean oil at 100 °C, a methanol/oil ratio of 20:1 in a continuous reactor with a mass flow of

250 g/h and a residence time of 5 h; the yield to FAME achieved was 75 %. Kutálek et al. (2014)

performed rapeseed oil transesterification in a continuous flow reactor and packed bed, using

Mg-Al-K and Mg-Al-Na mixed oxide catalyst, with a methanol/oil ratio of 24:1, at 140 °C and

residence time of 1.54 h, reaching a yield to FAME of 77 %. Likewise Ren et al. (2012) assessed

the behavior of an activated resin for the transesterification of soybean oil in a continuous flow

reactor using a methanol/oil ratio of 9:1, at 50 °C and a residence time of 56 min; a yield to

FAME of 95.2 % was obtained.

The aim of this work was to assess the catalytic performance of pelletized biomass fly ash in

the production of FAME in a continuous fixed bed reactor, using mixture of Refined Palm Oil

(RPO) and Waste Cooking Oil (WCO) with methanol as raw materials. An experimental Box-

Benhken design and Response Surface Methodology (RSM) was used for the optimization and

interpretation of the results obtained. Thus, this work contributes to decrease the

environmental burdens (and costs) of the conventional biodiesel process, by using residual

materials, in a continuous regime, about which there is some lack of knowledge in the research

literature.


90


The Box-Benhken experimental design and Response Surface Methodology (RSM) were used

to design the experiments to optimize the FAME production process in a continuous fixed bed

reactor and for the data treatment.

Previously, the pelletized solid catalyst (FAD) was prepared and characterized in terms of some

of its chemical, physical and structural properties. The raw-material for FAME synthesis

consisted of a blends of WCO and RPO in different ratios. The adopted procedures are

described in the next sections as well as the analytical methods used and design and

construction of continuous reactor.


Three mixtures were prepared using different percentages of RPO and WCO, namely:

M1 (100 wt% RPO), M2 (50 wt% RPO, 50 wt% WCO) and M3 (100 wt% WCO), which were

characterized in terms of acid value (NTC 218) (ICONTEC 218 1999), density (NTC 336)

(ICONTEC 1998), and saponification number, SN (NTC 335) (ICONTEC 1998). The SN was used

to calculate molecular weight (MW) using the Equation 4.1 (Mansir et al. 2018).

𝑀𝑊 =56,1×1000×3

𝑆𝑁 (4.1)

The FFA content was calculated from the acid value (AV, mgKOH/g) using the Equation 4.2

(Mansir et al. 2018).

FFA =𝐴𝑉

2 (4.2)


The catalyst pellets were prepared using the protocol described by Da Silva et al. (2014), using

biomass fly ash. This ash was collected in a dedusting flue gas equipment (electrostatic


91

precipitator) of a thermal power-plant using residual eucalyptus biomass as fuel, located in the

Centre Region of Portugal.

The particle size analysis of the fly ash (powder) showed the following size distribution (ASTM

series): 66.4 wt% has a particle size in the 200-230 mesh with diameters ranging between 75

μm and 63 μm, respectively; 33.6 wt% has a particle size smaller than 230 mesh. A mixture of

fly ash (60.94 wt%), soluble starch (1.2 wt%) and distilled water (37.86 wt%) were used for the

catalyst pellet manufacturing. The resulting aqueous paste was extruded using a syringe with

2 mm internal diameter tip, obtaining pellets 1 to 1.5 cm in length, which were dried in an oven

at 120 °C for 12 h.

Figure 4.1 shows pictures of biomass fly ash powder used in the pellets (Figure 4.1 a) and the

pellets ready (Figure 4.1 b) to be used for the FAME production.

(a) (b)

Figure 4.1- Biomass fly ash catalyst: (a) powder (before pelletization), (b) pelletized.

The pelletized biomass fly ash catalyst was further characterized in terms of: (i) crystallographic

structures using X-Ray Diffraction (XRD, PAN Analytical Empyrean X-ray diffractometer

equipped with Cu-Kα radiation source λ= 1.54178 Å at 45 kV/ 40 mA); (ii) surface area, pore

size and pore volume was carried out by nitrogen adsorption studies at 77 K using a surface

area analyzer Micromeritics Gemini V- 2380); (iii) surface morphology using surface scanning

electron microscopy (SEM, using Hitachi SU-70 microscope operating at 15 kV); (iv) functional


92

surface species by infrared Fourier transform (FTIR, Agilent CARY 630 with a wave number

range of 400 to 4000 cm−1); (v) superficial elemental atomic concentration by means of X-ray

photoelectronic spectroscopy (XPS, using a PHOIBOS 150 2D-DLD equipment, with a

monochromatic Al K-radiation source -FOCUS 500- operated at 100 W) and (vi) basic and acid

strength by using Hammett indicators following the protocol described by Vargas et al. (2019a)

(for basic strength: neutral red, pKa = 6.8; bromothymol blue, pKa = 7.2; phenolphthalein, pKa

= 9.3; indigo carmine, pKa = 12.2; and 2,4-dinitroaniline, pKa = 15.0; indicators for acid

strength: bromothymol blue, pKa = 7.2; neutral red, pKa = 6.8; bromocresol purple, pKa = 6.1;

bromocresol green, pKa = 4.7; and bromophenol blue, pKa = 3.8).

4.2.3 Experimental setup

Figure 4.2 and Figure 4.3 show the experimental setup assembled for this work. It had three

main zones: (1) feed, (2) reaction and (3) separation. In the feed zone there was a closed glass

container with a capacity of 1 L where the reagents (methanol and oil mixture) were mixed

and heated (heating plate). This mixture was pumped from the V-100 to the bottom of the

continuous fixed bed reactor (Figure 4.2. Reaction zone) using a peristaltic pump (Masterflex

and model: EW-77122) and a rotameter (Gilmond) was used to measure the flowrate.


93

V-100 – glass container with methanol and oil H-100 – heating plate P-100 – feed pump

F-100 – rotameter R-100 – fixed bed reactor TB-100 – thermostatic recirculating bath TT– thermocouple

V-101 – glass container C-100 – condenser SF-100 – separatory funnel V-102 – crude glycerin reservoir V-103 – FAME reservoir

Figure 4.2 - Experimental setup diagram used in this work (1. Feed zone, 2. Reaction zone, 3.

Separation zone).

For each oil mixture tested, a pump calibration was carried out to guarantee the residence

time (RT) set for each experiment. The operating flowrates used in this work ranged between

0.28 mL/min and 0.85 mL/min which were used together with the packing density and the

fixed bed void fraction to calculate the residence time of the reactor in each experiment

(Chattopadhyay and Ramkrishna 2013). The reactor was a glass tube with an internal diameter

of 21.9 mm and 320 mm height, with two stainless-steel mesh to support the catalyst.

The temperature inside the reactor was controlled by a heating jacket connected to a

thermostatic bath (Lauda Alpha A6). In order to ensure a uniform temperature inside the

reactor there was a thermocouple type K (Watlow AW) with three measuring points along the

catalytic fixed bed.


94

Figure 4.3 - Assembly of the fixed bed reaction system for continuous FAME production.

Downstream of the reactor there was the separation step (Figure 4.2. Separation zone), where

the reactor's outflow was separated into 3 fractions: methanol, crude glycerin and FAME. The

devices used for the separation were a glass container (V-101), a condenser and separatory

funnel. Firstly, the excess methanol was separated by heating the mixture of V-101 and

recovered the alcohol by condensation (C-100). Then, the remaining mixture (in V-101) was

transferred to a separatory funnel in order to separate the crude glycerin from FAME (light

phase). These two products were then stored in different reservoirs (V-102 and V-103). The

sampling was performed from FAME’s reservoir for chromatographic (GC) analysis.

4.2.4 FAME content

After the separation step, the collected samples were characterized in terms of FAME content

with a SHIMADZU G-C 2014 gas chromatograph (GC) equipped with a flame ionization detector

(FID), a capillary column SGEBP-20 (60 m x 0.25 mm ID x 0.25 μm) and using helium as a carrier

gas with a flow rate of 16.7 mL/min and a pressure of 2.5 atm. The injector (AOC-20i) was

1 2 3


95

operated at 200 oC and an injection volume of 2.0 µL in Split mode. Methyl heptadecanoate

was used as an internal standard and the samples were solubilized in hexane grade HPLC. The

samples to be analyzed were firstly dehydrated (using anhydrous sodium sulfate) and filtered

(using Thermo Scientific Nalgene membranes of 0,2 μm). Then, one mixed 500 μL of the phase

rich in esters (i.e., the solution resulting from the previous step) in hexane (10 mg / 10 mL),

200 μL of methyl heptadecanoate solution in hexane (10 mg / 10 mL) and 800 μL of hexane.

The FAME concentration was calculated using Equation 4.3 (AENOR-EN 14103 2011).


𝐴𝐸𝐼×

𝑊𝐸𝐼

𝑊 (4.3)

Where C is the concentration of FAME in the purified final mixture (w/w),∑ 𝐴 is the total peak


heptadecanoate, 𝑊𝐸𝐼 is the mass (mg) of methyl heptadecanoate used and is the mass (mg) of

the sample used in the analysis.

4.2.5 Preliminary assays for steady state identification

Some preliminary assays were carried out to identify the time at which the steady state of the

continuous fixed bed reactor used in this work was reached. The range of resident time,

WCO/RPO mass ratio and methanol/oil molar ratio tested in these assays were chosen based

on data reported by different authors and shown in Table 4.1. These assays were carried out

at temperature T = 60 °C with a catalytic bed depth of 15 cm (bulk density of 2.16 g/mL) around

14 h of continuous operation. No replicates of these assays were performed.


96

Table 4.1 - Operating conditions of continuous flow reactors to produce FAME.

Oil Catalyst T

(°C) RT

(min) Methanol/oil

(mol/mol) Reference

Sunflower Calcium oxide 60 120 12:1 (Miladinović et al. 2020)

WCO Sodium hydroxide 60 75 6:1 to 12:1 (Avellaneda and Salvadó 2011)

Soybean Resin D261 activated 50 56 9:1 (Ren et al. 2012)

Cottonseed Sodium silicate calcined

55 180 12:1 (Gui et al. 2016)

Soybean Sodium silicate 60 65 9:1 (Luo et al. 2017)

The results obtained in these assays are shown in Figure 4.4, where it is possible to identify the

steady state achievement at approximately 6h of operating time. Furthermore, it is observed

that a 3:1 molar ratio methanol/oil produces the lower content of FAME whereas the 12:1

molar ratio produces the higher one, and does not have a significant difference with respect

to 18:1 for the same RT. Therefore, the range from 6:1 to 12:1 (mol/mol) for methanol/oil was

selected in the optimization study (see next section). Regarding the residence times tested (60,

120 and 180 min), they were maintained in the process’ optimization study given the high

concentrations of FAME obtained in these preliminary assays.


97

Figure 4.4 - Preliminary assays to identify the steady state of the continuous fixed bed

reactor.

In short, for the residence time ranging between 60 and 180 min, 7, 8 and 9 h were set as

sampling times in the next experimental design, since the steady state is assured.

4.2.6 Experimental design for optimization of the FAME production

Software Design - Expert 11.1.0 was used to perform the processing and analysis of

experimental data. The experimental Box-Benhken design is a spherical and rotating design

(Mishra et al. 2008), which consists of the center point and midpoints of the edges. It is

represented as a shape consisting of three interlaced 22 factorial designs (Aslan and Cebeci

2007) and a center point. Likewise, the RSM creates correlations, evaluates the effects of

various factors and their interactions to obtain the system response (Liu et al. 2014;

Salamatinia et al. 2010).

The influence on FAME concentration (response variable) of three factors – residence time,

WCO/RPO mass ratio and methanol/oil molar ratio – was investigated. Table 4.2 presents the

experimental range and factor levels tested in this work.


98

Table 4.2 - Factor and levels of process variables used in the Box – Behnken experimental

design.

Real variables Coded variables Level

Low (-1) Medium (0) High (+1)

Residence time (min) A 60 120 180

WCO/RPO (wt%) B 0 (M1) 50 (M2) 100 (M3)

Methanol/oil (mol/mol) C 6 9 12

All experiments were carried out using a fixed bed depth of 15 cm (bulk density: 2.16 g/mL),

temperature of 60 °C and an operating time of 9 h. As stated before, sampling of purified FAME

(from reservoir V-102, Figure 4.2) was performed at 7, 8 and 9 h for GC characterization. The average

of these three points is taken as the value of the FAME concentration (experimental).

Seventeen experimental assays were conducted, including five repetitions of the center point. The

correlation in the form of a quadratic polynomial equation was developed to predict the response

(FAME concentration) based on independent variables (or factors) and their interactions according

to Equation 4.4.

𝑌 = 𝛽0 + ∑ 𝛽𝑖𝑥𝑖𝑘𝑖=1 + ∑ 𝛽𝑖𝑖𝑥𝑖

2𝑘𝑖=1 + ∑ ∑ 𝛽𝑖𝑗𝑥𝑖𝑥𝑗

𝑘𝑗=𝑖+1

𝑘𝑖=1 + 𝜀 (4.4)

Where 𝑌 is the predicted response for the process, i.e., the dependent variable; 𝛽0 is the

intercept coefficient (offset); 𝛽𝑖 are the linear terms; 𝛽𝑖𝑖 are the quadratic terms; 𝛽𝑖𝑗 are the

interaction terms; 𝑥𝑖 and 𝑥𝑗 are the independent variables; and 𝜀 is the error (Hajamini et al.

2016).

The inference in the regression model was made through an analysis of variance (ANOVA), for

a 95 % confidence level, where statistically significant factors in the response variable were

identified, and an analysis of the model determination coefficients, R2 and R2 adjusted

(Adj R2), were used to evaluate the degree of adjustment of the regression model to the

experimental data.

Validation of the assumptions of the regression model was performed through a residual

analysis (normality and residual plots). This analysis was based on standardized residues.


99

Once the best regression model was selected and validated, optimal reactor operating

conditions were identified using RSM. Then, for model validity purposes, three runs were

performed using the optimal conditions, which allowed to determine the deviations between

the data predicted by the model and the experimental ones. Additionally, the catalytic stability

of the biomass fly ash fixed bed was evaluated through an assay during 32 h of operation,

where the FAME concentration was monitored over time. The optimal operating conditions

found by the regression model were used in this assay.



The properties of the oil mixtures prepared for this study are shown in the Table 4.3.

Table 4.3 - Properties of the oil mixtures used.

M1 M2 M3

%WCO 0 50 100

%RPO 100 50 0

Density (g/mL) 0.896 ± 0.005 0.905 ± 0.006 0.910 ± 0.005

𝐴𝑉 (mg KOH/g) 0.052 ± 0.008 4.254 ± 0.062 7.328 ± 0.152

FFA (wt%) 0.026 ± 0.008 2.127 ± 0.032 3.664 ± 0.069

𝑀𝑊 (g/mol) 826.444 ± 0.252 837.189 ± 1.586 847.935 ± 2.681

The properties of M1 (RPO) are similar to those reported by Kansedo et al. (2009), Ñústez et

al. (2019) and Singh and Dipti (2010). Regarding the properties of the oil mixture M3, Wan

Omar et al. (2011a) and Lam et al. (2010) found results similar for both FFA (wt%) and AV (mg

KOH/g). Seeing the WCO has an FFA content less than 15 wt% it can be classified as yellow fat

(Avhad and Marchetti 2015).


100


The solid catalyst prepared was characterized for some textural properties such as specific

surface area, crystalline structure, surface functional groups, but also their basic and acid

strength, etc. The results are shown and discussed in the next sections.

4.3.2.1 XRD analysis

The XRD diffractogram of the biomass fly ash catalyst is shown in Figure 4.5 where there are

the diffraction peaks corresponding to sylvite phase (potassium chloride, KCl,ICDDPDF4+00-

041-1476) detected at 2 = 28.4°, 40.7° and 50.2°, calcite phase (calcium carbonate, CaCO3,

ICDDPDF4+01-080-9776) detected at 2 = 23.2°, 29.3°, 36.0°, 39.1°, 43.4°, 47.7°, 48.7° and

57.7°, portlandite phase (Calcium hydroxide, Ca (OH)2, ICDDPDF4+01-083-4600) detected at

2 = 47.7° and 64.9° and quartz phase (silicon dioxide, SiO2, ICDDPDFF4+04-006-1757)

detected at 2 = 21.0° and 26.7°. Similar results were reported by Miladinović et al. (2020) for

walnut ash; these authors found compounds of CaO, Ca(OH)2 and SiO2. Indeed, the combustion

of biomass residues generates compounds of CaO, Ca(OH)2 and CaCO3 as described by Osman

et al. (2018) and Luque et al. (2012). These alkaline species (which order of basicity is: oxide >

hydroxide > carbonate form) and the sylvite species (Betiku et al. 2016; Nath et al. 2019) are

able to catalyse the reaction of FAME production (Yoosuk et al. 2011).

Figure 4.5 - XRD Diffractogram of biomass fly ash catalyst (• SiO2, ■ CaCO3, ▲ KCl, ★ Ca

(OH)2).


101

The absence of CaO and the presence of Ca(OH)2 and CaCO3 in the pelletized biomass fly ash

(Figure 4.5) can be explained by hydration and carbonation reactions that occurred during the

preparation of this catalyst from powder ash. Indeed, the CaO existing in the powder ash

(characterized in a previous work (Vargas et al. 2019b)) must have been converted to Ca(OH)2

and CaCO3 due to the contact with water and atmospheric air.

4.3.2.2 Hammett indicators and BET surface area analyses

Biomass fly ash used in this work has an intermediate basic strength (10.1≤ pKa <12.2), due to

the high basicity of the metal-oxygen groups (Lewis bases) present in the calcium compounds

and alkaline metal (K and Na) on its surface (see Sections 4.3.2.1 and 4.3.2.5) and a low acid

strength (6.8≤ pKa<7.2) (Pavlović et al. 2020).

This ash has a surface area (BET) of 4.8568 m2/g, a pore volume of 0.009691 cm3/g and a an

average pore diameter of 74.853 Å, which is according to the reported data by Chakraborty et

al. (2010). Given the pore diameter, this material has favorable characteristics for the

absorption and desorption of triglyceride molecules, diglycerides, monoglycerides, glycerin

and FAME, since it is greater than 58 Å (Jacobson et al. 2008).

Figure 4.6 shows nitrogen absorption-desorption isotherm the type IV for the catalyst

assessed, which is in agreement with the classification proposed by the International Union of

Pure and Applied Chemistry (Thommes et al. 2015). The isotherms’ behavior is characteristic

of mesoporous catalysts and low surface area.

Figure 4.6 - Absorption and desorption isotherms of biomass fly ash.

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0.0 0.2 0.4 0.6 0.8 1.0

Qu

anti

ty a

dso

rbed

(cm

³/g)

Relative pressure (p/p°)


102

4.3.2.3 SEM analysis

Figure 4.7 shows the SEM images for the biomass fly ash catalyst.

(a) (b)

Figure 4.7 - SEM images of biomass fly ash using a magnification of 500X (a) and 5000X (b).

In Figure 4.7 a, the particles have irregular shapes of different sizes, this morphology is like that

found by Wang et al. (2008) for wood ashes. On the other hand, in Figure 4.7 b, the catalyst

particles have a diameter near 100 nm, which is similar to the eucalyptus fly ashes

characterized by Rajamma et al. (2009).

4.3.2.4 FTIR analysis

The FTIR spectrum of the biomass fly ash catalyst used in this work is shown in Figure 4.8. A

major absorption broadband is observed at 1401 cm-1 and minor absorption bands at 712, 872,

2855 and 2922 cm-1, corresponding to asymmetric stretching and vibration modes in the

carbonate group (CO3-2). This result confirms the presence of CaCO3 in this ash already

detected by XRD.


103

Figure 4.8 - FTIR spectrum of biomass fly ash catalyst.

Groups PO4-3 and Si-O (phosphates and silicon oxides) show wide absorption bands in the

region of 1096 cm-1; similar bands were found by Maneerung et al. (2015) and Liang et al.

(2020) in residues of fly ash from gasification of woody biomass and fly ash of rice husk

respectively.

4.3.2.5 XPS analysis

The biomass fly ash catalyst wide energy spectrum is shown in Figure 4.9. An intense peak of

O1s can be seen at 528.68 eV, and peaks at 1069.06, 343.65, 281.65, 194.88 and 166.60 eV for

Na1s, Ca2p, C1s, Cl2p and S2p, respectively. Yang et al. (2019) found similar binding energy

peaks in the spectrum for carbon ash samples.

Table 4.4 shows the wide scanning result of surface (layer up to 10 nm) chemical composition

of biomass fly ash. Elements O and C are presented in higher atomic concentrations on the

surface of the catalyst.

712

872

1096

1401

2855

2922

0

20

40

60

80

100

120

40080012001600200024002800320036004000

% T

ran

smit

ance

Wavenumber (cm-1)


104

Table 4.4 - Summary of XPS analysis done on the surface of biomass fly ash catalyst.

Sample

Atomic concentration (%)

C O Ca Mg Cl Na Si Al K S P

Biomass fly ash

31.16 35.48 13.95 5.05 2.18 0.75 0.81 0.68 7.62 1.56 0.76

Figure 4.9 - The XPS wide energy spectrum of biomass fly ash catalyst.

A significant amount of alkaline metals (Na and K) but mainly alkaline earth metals (Mg and

Ca) was registered, which give the basic character to this solid material. It was also recorded

the presence of Cl, Si, Al, S and P; some of these elements (e.g. Si and Al) give an acidic

character to the catalyst (Pavlović et al. 2020). Similar atomic concentrations were found by

Rajamma et al. (2009) for C, O, Ca and Na in fly ash of eucalyptus biomass of 38.7 %, 29.0 %,

12.1 % and 1.0 %, respectively.

4.3.3 Optimization of FAME production process: regression model and statistical analysis

The experimental results and those predicted by the regression model (Equation 4.5) for each

set of operating conditions (see Table 4.2) are shown in Table 4.5.


105

Table 4.5 - Experimental and predicted results of RSM.

Run

Real variables FAME concentration (wt%)

RT (min)

WCO/RPO (wt%)

Methanol/oil (mol/mol)

Experimental Predicted*

1 60 50 12 84.9 84.7

2 120 100 12 85.3 83.3

3 180 0 9 72.4 70.2

4 120 0 6 59.8 61.8

5 180 50 12 92.0 92.1

6 180 50 6 75.0 75.3

7 120 50 9 85.11 84.9

8 120 50 9 84.5 84.9

9 60 0 9 52.8 50.9

10 120 100 6 62.2 60.1

11 180 100 9 61.1 63.0

12 120 50 9 84.8 84.9

13 120 50 9 84.7 84.9

14 120 0 12 75.8 77.9

15 60 100 9 59.4 61.7

16 60 50 6 62.2 62.1

17 120 50 9 85.5 84.9

*Predicted by the regression model

The ANOVA results for the quadratic polynomial model are shown in Table 4.6, where one can

see that the model has a good fit with an R2 = 0.9859 and Adj R2 = 0.9678. The R2 value indicates

that the model predicts 98.6 % of the response variability. However the lack of fit is significant

(p-value 0.0006), which is undesirable. The lack-of-fit test is used as support test for adequacy

of the fitting model. A significant lack of fit means that the variation of the repeats around

their mean values is lower than the variation of the design points around their predicted values


106

(Miladinović et al. 2016). Nonetheless, in Figure 4.10 c one observe that predicted and

experimental values are similar, therefore the lack of fit may be due to the existence of a

systematic variation (experimental values) that cannot be explained by the regression model

(Miladinović et al. 2016). Moreover, as will be discussed later, the experimental FAME

concentration registered under the optimal operating conditions, obtained from the model,

was close to the one predicted by the model.

Table 4.6 - ANOVA table of the regression model.

Source of variations

Sum of squares

Degrees of

freedom

Mean square

F - value p – value

Model 2442.88 9 271.43 54.51 < 0.0001

Residual 34.85 7 4.98

Lack of fit 34.23 3 11.41 73.78 0.0006

Pure error 0.6187 4 0.1547

Total 2477.73 16

R2 = 0.9859 Adj Pred C.V.a = 2.99 % S.D.b = 2.23

R2 = 0.9678 R2 = 0.7785

a C.V.= coefficient of variation. b S.D.= standard deviation.

The high F-value (54.51) and the low p-value (<0.0001) of the model means that it is statistically

significant. The "Pred R2" and the "Adj R2" values are consistent since their difference is less

than 0.2 (note reported by Design Expert software); which confirm that the fitting regression

model satisfactorily predicts the effect of the three factors evaluated on FAME concentration.

Equation 4.5 represents the fitting quadratic polynomial model (coded variables).

𝑌 = 84.93 + 5.16 𝑥𝐴 + 0.9093 𝑥𝐵 + 9.82 𝑥𝐶 − 4.49 𝑥𝐴 × 𝑥𝐵 − 1.45 𝑥𝐴 × 𝑥𝐶 + 1.78 𝑥𝐵 × 𝑥𝐶

−7.78 𝑥𝐴2 − 15.64 𝑥𝐵

2 + 1.49𝑥𝐶2 (4.5)

Where 𝑌 is the response variable (FAME concentration, wt%), 𝑥𝐴 (residence time, min), 𝑥𝐵

(WCO/RPO, wt%) and 𝑥𝐶 (methanol/oil, molar ratio) are the factors studied. The positive sign

of a coefficient means synergistic effect while the negative sign shows an opposite effect on

variables that influence the FAME concentration (Liu et al. 2014).


107

The statistical significance of each regression coefficient of the model in the response variable

was evaluated and the results are shown in the Table 4.7. The terms 𝑥𝐴, 𝑥𝐶, 𝑥𝐴 × 𝑥𝐵, 𝑥𝐴2 and

𝑥𝐵2 are significant since the respective p-values are lower 0.05 (Avramović et al. 2010). Besides

that, the influence of square value of residence time (𝑥𝐴2) and WCO/RPO (𝑥𝐵

2) have a significant

negative effect (on the FAME concentration) of -7.87 (p-value = 0.0002) and -15.64

(p-value < 0.0001), respectively. The linear term of the molar methanol/oil ratio (𝑥𝐶) has

greatest positive effect of 9.82 (p-value < 0.0001) on the FAME concentration. The interaction

terms 𝑥𝐴 × 𝑥𝐶 (p-value = 0.2355) and 𝑥𝐵 × 𝑥𝐶 (p-value = 0.1546) are not statistically

significant. On the other hand, 𝑥𝐴 × 𝑥𝐵 (p-value = 0.005) is statistically significant, which

means that there is interaction between the variables RT and WCO/RPO ratio (wt%). In other

words, the influence of WCO/RPO ratio (wt%) on the FAME concentration depends on the RT,

which could be due to viscosity effects on resistance to mass transfer (the higher WCO/RPO

ratio (wt%), the higher viscosity (Vargas et al. 2019a)) and to the reversibility of esterification

and transesterification reactions.

Table 4.7 - Regression coefficients (coded factors) for the fitting quadratic polynomial model.

Model parameters Estimate coefficient Standard error p - value

Intercept 84.93 0.99

𝑥𝐴 5.16 0.79 0.0003

𝑥𝐵 0.9093 0.79 0.2869

𝑥𝐶 9.82 0.79 < 0.0001

𝑥𝐴 × 𝑥𝐵 -4.49 1.12 0.0050

𝑥𝐴 × 𝑥𝐶 -1.45 1.12 0.2355

𝑥𝐵 × 𝑥𝐶 1.78 1.12 0.1546

𝑥𝐴2 -7.87 1.09 0.0002

𝑥𝐵2 -15.64 1.09 < 0.0001

𝑥𝐶2 1.49 1.09 0.2136

In summary, all the variables studied have a statistically significant effect on FAME

concentration, but not in the same way. Concerning the residence time both linear and


108

quadratic terms are significant, on the other hand, only the quadratic term of the WCO/RPO

ratio (wt%) affects (negatively) the response variable. With regard to the molar ratio

methanol/oil only the linear term is statistically significant. There is no interaction between

the independent variables, except for the residence time and WCO/RPO ratio (wt%).

The residual plots shown in Figure 4.10 were used to validate the model assumptions.

Figure 4.10 a shows a normal probability plot of residues, which corresponds to the difference

between the experimental response and the predicted by the model. The data are along a

straight line and without the distribution presenting S-shape, so it can be concluded that the

residues follow a normal distribution (Hajamini et al. 2016; Jaliliannosrati et al. 2013).

The graph of the residuals vs. the predicted values is illustrated in Figure 4.10 b and shows the

data randomly distributed with equal number of points above and below the central horizontal

line and between two horizontal lines at a threshold of ± 3.00 (three standard deviations)

(Noshadi et al. 2012).

The experimental values of FAME concentration vs. the values predicted by the model are

shown in Figure 4.10 c, which corroborates the tuning effectiveness of the developed

regression model. In short, this analysis confirms the accuracy and reliability of the proposed

regression model.


109

(a) (b)

(c)


(c) Predicted versus experimental values plot.

Figure 4.11 shows the response surface graphs for three evaluated variables.


110

(a) (b)

(c)

Figure 4.11 - FAME concentration response surface graphics based on: (a) Residence time and

WCO/RPO ratio (wt%) at molar methanol/oil ratio of 9 mol/mol; (b) Time of

residence and molar methanol/oil ratio at WCO/RPO ratio of 50 wt%; (c) Molar

methanol/oil ratio and WCO/RPO ratio (wt%) at time of residence of 120 min.

Residence time up to 120 min favors significantly the FAME concentration

(see Figure 4.11 a&b); for RT greater than 120 min, the FAME concentration does not rise,

regardless of the WCO/RPO (wt%) used (Figure 4.11 a) and, depending on methanol/oil molar

ratio (Figure 4.11 b), may even slightly decrease. The reversibility of (global) reaction of FAME

production may be one of the reasons for this. Gui et al. (2016) observed similar trends in a

residence time range between 60 and 180 min using sunflower oil and charred sodium silicate

as catalyst.

0.0

20.0

40.0

60.0

80.0

100.0

60

80

100

120

140

160

180

020

4060

80

FA

ME

co

nce

ntr

atio

n (

wt%

)

Res

iden

ce ti

me

(min

)

WCO/RPO (wt%)

0

0-20 %

20-40 %

40-60 %

60-80 %

80-100 %

0.0

20.0

40.0

60.0

80.0

100.0

60

80

100

120

140

160

180

67

89

1011

FA

ME

co

nce

ntr

atio

n (

wt%

)

Res

iden

ce ti

me

(min

)

Methanol/Oil (mol/mol)

0

0-20 %

20-40 %

40-60 %

60-80 %

80-100 %

0.0

20.0

40.0

60.0

80.0

100.0

0

20

40

60

80

100

67

89

1011

FA

ME

co

nce

ntr

atio

n (

wt%

)

WC

O/R

PO

(wt%

)

Methanol/Oil (mol/mol)

0

0-20 %

20-40 %

40-60 %

60-80 %

80-100 %


111

Concerning the effect of WCO/RPO (wt%), Figure 4.11 a&c show that for the range between

50 and 80 wt% the highest FAME concentrations are achieved. This means that till certain

percentages in the oil mixture, the WCO promotes higher FAME concentrations which, in turn,

is a promising outcome for the economic and environmental sustainability of the process

(Vargas et al. 2019b). Moreover, Figure 4.11 b&c show the great influence of the methanol/oil

molar ratio on the response variable; 12:1 value reached the highest FAME concentration

around 92 %. Similar results were found by Ni and Meunier (2007), who observed the higher

FAME concentrations for methanol/oil ratio of 12:1; above this ratio did not present a

noticeable increase on FAME concentration.

The high concentrations of FAME registered in this work can be due to the majority presence

of crystalline phases of calcium carbonate, calcium hydroxide and alkaline metals that bring

basic character to the catalyst (Yoosuk et al. 2011), which was observed by XRD and Hammett's

indicators. Thus, even in the absence of the most basic form of calcium (CaO), the species

present (Ca(OH)2 and CaCO3) and KCl have been shown to provide sufficient basicity to catalyze

efficiently the transesterification reaction. In addition, high concentrations of Ca (13.95 %), K

(7.62 %) and Mg (5.05 %) were found on its surface and pore diameters of 74.853 Å in the XPS

and BET, respectively, which are suitable for transesterification reactions.

The bifunctional character of the powder fly ash used to prepare the pelletized catalyst was

verified in a previous work (Vargas et al. 2019a). This catalyst feature can also explain the high

concentrations of FAME observed with oil mixtures with high percentages of WCO, i.e. with

high acid values. Actually, despite of its low acid strength, this catalyst seems to be able to also

catalyze the esterification reaction. According to some authors (Rabiah Nizah et al. 2014; Wan

Omar et al. 2011b) the balance of acid and basic catalyst sites plays an important role in the

bifunctional performance of the catalysts.

Optimal operating condition

An important objective of this research was to find the optimal operating conditions to achieve

the maximum FAME concentration, combining the various independent variables studied

using a regression model. The model showed two possible combinations of variables to achieve


112

the highest FAME concentration: (i) residence time of 124 min, WCO/RPO of 74.6 wt% and

molar methanol/oil ratio of 12:1, reaching a FAME concentration of 93.8 %; and (ii) residence

time of 135 min, WCO/RPO of 57.2 wt% and molar methanol/oil ratio of 12:1 reaching a FAME

concentration of 96.4 %.

The first set of operating conditions was chosen, since, it has the largest amount of WCO and

the least residence time, which results in a decrease in costs in raw materials and operating

times compared to the second set.

Aiming to check the optimal operating conditions found by the regression model, three

experiments were performed under the selected conditions shown in Table 4.8.

Table 4.8 - Experimental data on optimal operating conditions.

Run FAME concentration (wt%)

5 h 6 h 7 h Average S.D.

1 90.2 89.9 89.2 89.7 0.5

2 88.8 88.5 91.4 89.5 1.5

3 90.5 89.9 88.9 89.8 0.7

Experimental FAME concentration = 89.7 wt%

Predicted FAME concentration = 93.8 wt%

% error = 4.6 %

The average experimental FAME concentration was 89.7 % (± 0.8%), which is close to the

predicted by the regression model (93.81 %); with a 4. 6% error (± 0.05%) confirming the fit of

the found model.

4.3.4 Catalytic stability of the pelletized biomass ash

Figure 4.12 shows the results of the experiment, carried out under optimal operating

conditions, to assess the catalytic stability of the fixed bed over 32 h of continuous operation.

After steady state is reached, FAME concentration remains at c.a. 88 % until (at least) around

20 h. However, about 30 h there is a slight decrease ( 5.6%) in the FAME concentration, which


113

may indicate some loss of catalytic activity in the fixed bed. Other authors such as Ren et al.

(2012) and Da Silva et al. (2014) also checked the catalyst stability of other catalysts used for

FAME production. The first authors observed a 74.4 % loss of catalytic activity of D261 resin

after 8 h of operation, on the other hand, Da Silva et al. (2014) reported a quite small loss of

activity (3%) for a zinc oxide catalyst after 120 h of operation. In a perspective of scaling up the

process, the results of this work regarding the biomass fly ash pelletized catalyst performance

are promising and encouraging.

Figure 4.12 - FAME concentration of the catalyst over 32 h of continuous operation.

4.4 Conclusions

The continuous fixed bed reaction system designed and built for this work allowed to evaluate

the catalytic activity and stability of the pelletized residual catalyst in a simple and repeatable

way.

The evaluation of a pelletized residual catalyst of fly ash to produce FAME was performed,

using RPO and WCO mixtures and using RSM and an experimental Box Behnken design to

optimize the response variable (FAME concentration). A quadratic regression model was well

0.0

20.0

40.0

60.0

80.0

100.0

0 5 10 15 20 25 30 35

FAM

E co

nce

ntr

atio

n (w

t%)

Operating time (hours)


114

fitted to the experimental results (p-value < 0.0001), predicting as optimal conditions a

residence time of 124 min, a WCO of 74.6 wt% and a molar methanol/oil ratio of 12:1 to obtain

a FAME concentration of 93.8 %. Thus, the biomass fly ash has shown to have a bifunctional

character, i.e., capable of catalyzing the transesterification and esterification reactions, as high

concentrations of FAME have been registered for high WCO/RPO ratios.

In the tested range, the most statistically significant variable (with a 95 % confidence level)

which affected the FAME concentration was the methanol/oil ratio with a p-value <0.0001,

followed by the residence time (p-value = 0.0003).

High WCO/RPO percentages up to 74.6 % can be used to achieve high FAME concentration.

Thus, significant quantities of WCO can be valorized in the FAME production process.

The selected regression model had a good adjustment of the experimental results found with

an R2 = 0.9859 and Adj R2 = 0.9678; Furthermore, the precision and reliability of the regression

model were confirmed.

Three tests were conducted under optimal operating conditions, where average experimental

FAME concentration was 89.7 % (± 0.8%), which is close to the predicted value by the model

(93.8 %). The error found between statistical correlation and experimental data was 4.6 % (±

0.05 %). In addition, it was demonstrated that the pelletized residual catalyst can be used for

up to 32 h of operation without significant loss of its catalytic activity. In future work, it would

be important to extend the operating time until find a significant drop in catalytic activity. In

addition, foreseeing a future commercialization, the FAME properties must be characterized

in order to assess whether it meets the ASTM D-6751, EN 14214 or other standards.

This work is contributing to enhance knowledge on continuous process of biodiesel production

based on residual materials, and also to increase its competitiveness in relation to the

conventional process, widely installed at industrial scale.

In short, it is important to make the most of residual raw materials such as biomass fly ash or

WCO, as they represent a contribution on strengthening the circular economy by making good


115

use of these wastes for the energy generation (biodiesel). Thus, this research offers a

sustainable and affordable approach by minimizing the environmental impact and costs

generated by biodiesel production.

Acknowledgments

The authors thank to the Universidad Jorge Tadeo Lozano (Direction of Investigation, Creation

and Extension) for the financial assistance of this work. The authors thanks to FCT/MCTES for

the financial support to CESAM (UIDP/50017/2020 & UIDB/50017/2020), through national

funds. This work was also developed within the scope of CICECO (UIDB/50011/2020 &

UIDP/50011/2020), financed by national funds through the FCT/MCTES and when appropriate

co-financed by ERDF under the PT2020 Partnership Agreement. Márcia C. Neves acknowledges

FCT, I.P. for the research contract CEECIND/00383/2017 under the CEEC Individual 2017.

References


Andrigo, P, R Bagatin, and G Pagani. 1999. “Fixed Bed Reactors (Chapter).Pdf.” Catalysis Today 52: 197–221.

Aslan, N., and Y. Cebeci. 2007. “Application of Box-Behnken Design and Response Surface Methodology for Modeling of Some Turkish Coals.” Fuel 86(1–2): 90–97.

Avellaneda, Fredy, and Joan Salvadó. 2011. “Continuous Transesterification of Biodiesel in a Helicoidal Reactor Using Recycled Oil.” Fuel Processing Technology 92(1): 83–91.


Avramović, Jelena M. et al. 2010. “The Optimization of the Ultrasound-Assisted Base-Catalyzed Sunflower Oil Methanolysis by a Full Factorial Design.” Fuel Processing Technology 91(11): 1551–57.

Betiku, Eriola, Aramide Mistura Akintunde, and Tunde Victor Ojumu. 2016. “Banana Peels as a Biobase Catalyst for Fatty Acid Methyl Esters Production Using Napoleon’s Plume (Bauhinia Monandra) Seed Oil: A Process Parameters Optimization Study.” Energy 103:


116

797–806. http://dx.doi.org/10.1016/j.energy.2016.02.138.

Chakraborty, R., S. Bepari, and A. Banerjee. 2010. “Transesterification of Soybean Oil Catalyzed by Fly Ash and Egg Shell Derived Solid Catalysts.” Chemical Engineering Journal 165(3): 798–805. http://dx.doi.org/10.1016/j.cej.2010.10.019.

Chattopadhyay, Soham, and Ramkrishna Sen. 2013. “Development of a Novel Integrated Continuous Reactor System for Biocatalytic Production of Biodiesel.” Bioresource Technology 147: 395–400. http://dx.doi.org/10.1016/j.biortech.2013.08.023.

Evangelista-Flores, A. Alcantar-Gonzalez, F. S. Ramırez de Arellano Aburto, N. Cohen Barki, A. Robledo-Perez, J. M. Cruz-Gomez 2012). “Diseño de Un Proceso Continuo de Producción de Biodiedel.” Revista Mexicana de Ingeniería Química 11(1): 23–43. https://www.redalyc.org/articulo.oa?id=62031508011

Fang, Dong, Jinming Yang, and Changmei Jiao. 2011. “Dicationic Ionic Liquids as Environmentally Benign Catalysts for Biodiesel Synthesis.” ACS Catalysis 1(1): 42–47.

Freedman, B., E.H. Pryde, and T.L. Mounts. 1981. “Hour Screening Test for Alternate Fuels in Energy Notes for , Variables Affecting the Yields of Fatty Esters from Transesterified Vegetable Oils 1.” American Society of Agricultural Engineers 2(10): 385–90.

Gui, Xia, Sichen Chen, and Zhi Yun. 2016. “Continuous Production of Biodiesel from Cottonseed Oil and Methanol Using a Column Reactor Packed with Calcined Sodium Silicate Base Catalyst.” Chinese Journal of Chemical Engineering 24(4): 499–505. http://dx.doi.org/10.1016/j.cjche.2015.11.006.





Jaliliannosrati, Hamidreza, Nor Aishah Saidina Amin, Amin Talebian-Kiakalaieh, and Iman Noshadi. 2013. “Microwave Assisted Biodiesel Production from Jatropha Curcas L. Seed by Two-Step in Situ Process: Optimization Using Response Surface Methodology.” Bioresource Technology 136: 565–73. http://dx.doi.org/10.1016/j.biortech.2013.02.078.


117

Jensen, R. R., S. S. Brake, and J. M. Mattox. 2004. “Trace Element Uptake in Plants Grown on Fly Ash Amended Soils.” Toxicological and Environmental Chemistry 86(1–4): 217–28.


Kutálek, Petr, Libor Čapek, Lucie Smoláková, and David Kubička. 2014. “Aspects of Mg-Al Mixed Oxide Activity in Transesterification of Rapeseed Oil in a Fixed-Bed Reactor.” Fuel Processing Technology 122: 176–81.


Liang, Guangbing et al. 2020. “Production of Biosilica Nanoparticles from Biomass Power Plant Fly Ash.” Waste Management 105: 8–17. https://doi.org/10.1016/j.wasman.2020.01.033.


Luo, Qingliang et al. 2017. “Continuous Transesterification to Produce Biodiesel under HTCC/Na2SiO3/NWF Composite Catalytic Membrane in Flow-through Membrane Reactor.” Fuel 197: 51–57.

Luque, Rafael et al. 2012. “Carbonaceous Residues from Biomass Gasification as Catalysts for Biodiesel Production.” Journal of Natural Gas Chemistry 21(3): 246–50.



Miladinović, Marija R. et al. 2016. “Modeling and Optimization of Sunflower Oil Methanolysis over Quicklime Bits in a Packed Bed Tubular Reactor Using the Response Surface Methodology.” Energy Conversion and Management 130: 25–33.

Miladinović, Marija R. et al. 2020. “Valorization of Walnut Shell Ash as a Catalyst for Biodiesel Production.” Renewable Energy 147: 1033–43.


118

Miller, B., D. Dugwell, and R. Kandiyoti. 2006. “The Fate of Trace Elements during the Co-Combustion of Wood-Bark with Waste.” Energy and Fuels 20(2): 520–31.

Mishra, Abha, Sunil Kumar, and Sudhir Kumar. 2008. “Application of Box-Benhken Experimental Design for Optimization of Laccase Production by Coriolus Versicolor MTCC138 in Solid-State Fermentation.” Journal of Scientific and Industrial Research 67(12): 1098–1107.

Mittelbach, Martin, Manfred Wörgetter, Josef Pernkopf, and Hans Junek. 1983. “Diesel Fuel Derived from Vegetable Oils: Preparation and Use of Rape Oil Methyl Ester.” Energy in Agriculture 2(C): 369–84.

Nath, Biswajit, Bipul Das, Pranjal Kalita, and Sanjay Basumatary. 2019. “Waste to Value Addition: Utilization of Waste Brassica Nigra Plant Derived Novel Green Heterogeneous Base Catalyst for Effective Synthesis of Biodiesel.” Journal of Cleaner Production 239: 118112. https://doi.org/10.1016/j.jclepro.2019.118112.

Ni, J., and F. C. Meunier. 2007. “Esterification of Free Fatty Acids in Sunflower Oil over Solid Acid Catalysts Using Batch and Fixed Bed-Reactors.” Applied Catalysis A: General 333(1): 122–30.

Noshadi, I., N. A.S. Amin, and Richard S. Parnas. 2012. “Continuous Production of Biodiesel from Waste Cooking Oil in a Reactive Distillation Column Catalyzed by Solid Heteropolyacid: Optimization Using Response Surface Methodology (RSM).” Fuel 94: 156–64. http://dx.doi.org/10.1016/j.fuel.2011.10.018.

Ñústez-Castaño, Stephanie Alexa, Duvan Oswaldo Villamizar-Castro, and Edgar Mauricio Vargas-Solano. 2019. “Evaluation of Dolomite as Catalyst in the Transesterification Reaction Using Palm Oil (RBD).” DYNA (Colombia) 86(209): 180–87.

Osman, Ahmed I., Abdelkader T. Ahmed, Christopher R. Johnston, and David W. Rooney. 2018. “Physicochemical Characterization of Miscanthus and Its Application in Heavy Metals Removal from Wastewaters.” Environmental Progress and Sustainable Energy 37(3): 1058–67.

Park, Young Moo et al. 2008. “The Heterogeneous Catalyst System for the Continuous Conversion of Free Fatty Acids in Used Vegetable Oils for the Production of Biodiesel.” Catalysis Today 131(1–4): 238–43.

Pavlović, Stefan M. et al. 2020. “A CaO/Zeolite-Based Catalyst Obtained from Waste Chicken Eggshell and Coal Fly Ash for Biodiesel Production.” Fuel 267(September 2019): 117171. https://doi.org/10.1016/j.fuel.2020.117171.

Pirez, Cyril et al. 2012. “Tunable KIT-6 Mesoporous Sulfonic Acid Catalysts for Fatty Acid Esterification.” ACS Catalysis 2(8): 1607–14.

Rabiah Nizah, M. F. et al. 2014. “Production of Biodiesel from Non-Edible Jatropha Curcas Oil via Transesterification Using Bi2O3-La2O3 Catalyst.” Energy Conversion and Management 88: 1257–62. http://dx.doi.org/10.1016/j.enconman.2014.02.072.


119

Rajamma, Rejini et al. 2009. “Characterisation and Use of Biomass Fly Ash in Cement-Based Materials.” Journal of Hazardous Materials 172(2–3): 1049–60.

Ren, Yanbiao et al. 2012. “Continuous Biodiesel Production in a Fixed Bed Reactor Packed with Anion-Exchange Resin as Heterogeneous Catalyst.” Bioresource Technology 113: 19–22. http://dx.doi.org/10.1016/j.biortech.2011.10.103.



da Silva, Fábio M. et al. 2014. “Continuous Biodiesel Production Using a Fixed-Bed Lewis-Based Catalytic System.” Chemical Engineering Research and Design 92(8): 1463–69. http://dx.doi.org/10.1016/j.cherd.2014.04.024.

Singh, S. P., and Dipti Singh. 2010. “Biodiesel Production through the Use of Different Sources and Characterization of Oils and Their Esters as the Substitute of Diesel: A Review.” Renewable and Sustainable Energy Reviews 14(1): 200–216.

Suarez, Paulo A.Z., Andre L.F. Santos, Juliana P. Rodrigues, and Melquizedeque B. Alves. 2009. “Biocombustiveis a Partir de Oleos e Gorduras: Desafios Tecnologicos Para Viabiliza-Los.” Quimica Nova 32(3): 768–75.

Tarelho, L. A.C. et al. 2015. “Characteristics of Distinct Ash Flows in a Biomass Thermal Power Plant with Bubbling Fluidised Bed Combustor.” Energy 90: 387–402. http://dx.doi.org/10.1016/j.energy.2015.07.036.

Teixeira, E. R. et al. 2019. “Recycling of Biomass and Coal Fly Ash as Cement Replacement Material and Its Effect on Hydration and Carbonation of Concrete.” Waste Management 94: 39–48. https://doi.org/10.1016/j.wasman.2019.05.044.

Thommes, Matthias et al. 2015. “IUPAC Technical Report Physisorption of Gases, with Special Reference to the Evaluation of Surface Area and Pore Size Distribution (IUPAC Technical Report).” https://www.3p-instruments.com/wp-content/uploads/2017/04/2015-IUPAC-Technical-Report.pdf (February 18, 2019).

Tran, Dang Thuan, Jo Shu Chang, and Duu Jong Lee. 2017. “Recent Insights into Continuous-Flow Biodiesel Production via Catalytic and Non-Catalytic Transesterification Processes.” Applied Energy 185: 376–409. http://dx.doi.org/10.1016/j.apenergy.2016.11.006.



120

Vargas, E. M., J. L. Ospina, L. A.C. Tarelho, and M. I. Nunes. 2019b. “FAME Production from Residual Materials: Optimization of the Process by Box–Behnken Model.” Energy Reports 6: 347–52. https://doi.org/10.1016/j.egyr.2019.08.071.

Vargas, Edgar M., Márcia C. Neves, Luís A.C. Tarelho, and Maria I. Nunes. 2019a. “Solid Catalysts Obtained from Wastes for FAME Production Using Mixtures of Refined Palm Oil and Waste Cooking Oils.” Renewable Energy 136: 873–83.

Vyas, Amish P., Jaswant L. Verma, and N. Subrahmanyam. 2010. “A Review on FAME Production Processes.” Fuel 89(1): 1–9. http://dx.doi.org/10.1016/j.fuel.2009.08.014.


Wan Omar, Wan Nor Nadyaini, and Nor Aishah Saidina Amin. 2011b. “Optimization of Heterogeneous Biodiesel Production from Waste Cooking Palm Oil via Response Surface Methodology.” Biomass and Bioenergy 35(3): 1329–38. http://dx.doi.org/10.1016/j.biombioe.2010.12.049.

Wang, Shuangzhen, Larry Baxter, and Fernando Fonseca. 2008. “Biomass Fly Ash in Concrete: SEM, EDX and ESEM Analysis.” Fuel 87(3): 372–79.

Yang, Lu et al. 2019. “Effect of the Intensification of Preconditioning on the Separation of Unburned Carbon from Coal Fly Ash.” Fuel 242(June 2018): 174–83.

Yoosuk, Boonyawan, Parncheewa Udomsap, and Buppa Puttasawat. 2011. “Hydration–Dehydration Technique for Property and Activity Improvement of Calcined Natural Dolomite in Heterogeneous Biodiesel Production: Structural Transformation Aspect.” Applied Catalysis A: General 395(1–2): 87–94. http://linkinghub.elsevier.com/retrieve/pii/S0926860X11000421 (May 29, 2015).

Zhang, Y., M. A. Dubé, D. D. McLean, and M. Kates. 2003. “Biodiesel Production from Waste Cooking Oil: 1. Process Design and Technological Assessment.” Bioresource Technology 89(1): 1–16.

121

SECTION E – Final remarks

Section E presents the main conclusions that can be drawn from this study, as well as some

limitations. It is also provided a list of suggestions for future works that can be developed in

the research field of this thesis.


122

FINAL REMARKS

123

5 Final remarks

5.1 General conclusions

This research work prepared, from waste materials, efficient bifunctional solid catalysts to

produce FAME from mixtures of low cost vegetable oils (WCO and RPO) with methanol. Among

four materials used, the fly ash of biomass dried exhibited the best performance in the catalysis

of both esterification and transesterification reactions. In this work it was also optimized the

production process of FAME in both batch reactor and in a continuous fixed bed reactor, using

biomass fly ash as (bifunctional) catalyst. Thus, through the use of waste materials it may be

possible to reduce the production costs of biodiesel and simultaneously prolong the life-cycle

of the materials in the economy, i.e., promoting a circular economy. Therefore, an awareness

should be created so that any material that is deemed a waste could be exploited for usage in

this or other applications.

It is important to mention that in the several experimental works performed in this thesis,

exhaustive characterization of the final product was not carried out, in order, for example, to

check if it meets any of the standards for the commercialization of this biofuel (biodiesel). Thus,

throughout this dissertation it was decided to designate the final product as FAME and not

biodiesel, since the "biodiesel" lable is reserved for the FAME product that meets a marketing

standard. However, the word biodiesel appears in the title of the thesis, and not the acronym

FAME, as it is the word of greatest public domain.

Concerning the main conclusions that can be drawn from these studies, they are summarized

in the following paragraphs.

Efficient heterogeneous catalysts were successfully prepared from solid waste materials for

FAME production by transesterification and esterification, using mixtures of refined palm oil

and waste cooking oil in different ratios and methanol (in batch reactor). The results


124

demonstrated that all the solid materials prepared and evaluated had different catalytic

performances. The best catalyst for catalyzing simultaneous both transesterification and

esterification reactions, i.e., having a bifunctional character (the balance of acid and basic

catalyst), was biomass fly ash dried (FAD), achieving yields and conversions above 95 % for a

blend of up to 25 wt% of WCO at 60 oC, 9:1 (mol/mol) of methanol to oil, 10 %wt catalyst

loading and over 180 min. The catalysts produced from dolomite rock (Dolomite C and

dolomite CSC) showed good performances of catalytic activity for transesterification (high

yields to FAME up to 87.8 %) and esterification (high conversion of FFA up to 100 %) reactions,

respectively. In fact, the sulfonation of Dolomite C was aimed at increasing its acid strength to

improve the catalytic activity for the esterification reaction. However, this treatment strongly

affected its ability towards the transesterification reaction, practically canceling it (yield to

FAME less than 3.4 %). While for the catalysts prepared from eggshells, the sulfonation

improved its ability for catalyzing the transesterification reaction, however the maximum

values attained did not exceed 70 % of FAME yield. On the other hand, this treatment

worsened the performance of this catalyst in FFA conversion (average conversion up to 62.7

%). Regarding the catalysts produced from PET, the results showed that they are good

candidates for catalyzing the esterification reaction of high acid value feedstocks reaching

FAME yield up to 88.9 %.

It is important to highlight that biomass fly ash dried (FAD) can be used as a catalyst as it

collected in the electrostatic precipitator equipment (directly and immediately), as its moisture

content is very low, with subsequent economic benefits.

The FAD underwent an optimization process using the RSM and a Box Behnken type

experimental design; testing it with methanol and mixtures of RPO and WCO in a batch reactor.

The solid catalyst showed a maximum FAME yield of 73.8 %, which can be achieved by using

13.57 wt% of catalyst loading, 6.7 methanol/oil molar ratio, 28.04 wt% of RPO/WCO mass ratio

and 55 oC for the reaction temperature (optimal operating conditions found). In the tested

range, the most significant variables were the RPO/WCO mass ratio and the reaction

temperature, followed by the catalyst loading. On the other hand, the methanol/oil molar ratio

FINAL REMARKS

125

was not significant; furthermore, biomass fly ash in powder can be used for up to three cycles

(batch mode) without loss of catalytic activity. However, the catalyst should be regenerated

between each cycle, by washing with isopropyl alcohol and calcined at 700 oC for 3 h. The

characterization of the surface, textural and crystalline properties of the catalyst, after use in

each FAME synthesis cycle, showed that those properties were not significantly affected. The

acid and basic strength remained constants.

Testing the FAD catalyst in a continuous reaction system is novel when using a fixed bed

reactor. Therefore, the powdered catalyst (FAD) was pelletized and a fixed-bed continuous

reaction system designed and constructed to perform a set of assays for the optimization of

FAME production. The regression model predicted as optimal conditions: a residence time of

124 min, a WCO of 74.6 wt% and a molar methanol/oil ratio of 12:1 to obtain a FAME

concentration of 93.8 %. Thus, the pelletized FAD showed to have again a bifunctional

character, that is, capable of catalyzing the transesterification and esterification reactions,

since high concentrations of FAME have been registered for high WCO/RPO ratios.

In the tested range, the most statistically significant variable (with a 95 % confidence level)

which affected the FAME concentration was the methanol/oil ratio followed by the residence

time. Moreover, high WCO/RPO percentages up to 74.6 % can be used to achieve high FAME

concentration. Thus, significant quantities of WCO can be valorized in the FAME production

process. It is important to mention that three assays were conducted under optimal operating

conditions, where average experimental FAME concentration was 89.7 % (± 0.8%), which is

close to the predicted value by the model (93.8 %). The error found between statistical

correlation and experimental data was 4.6 % (± 0.05 %). In addition, it was demonstrated that

the pelletized residual catalyst can be used for up to 32 h of operation without significant loss

of its catalytic activity.

Note: in principle, the drying step of the raw biomass fly ashes to prepare it for pelletization

could be skipped. In this work it was not done because one wanted to have the same (initial)

pretreatment procedure for all studies.


126

In this work it was found that biomass fly ash can be used as a solid catalyst in the production

of FAME just as it comes out of electrostatic precipitators, without the need of further

treatment, reaching high yields and with bifunctional characteristics. This means that it is

possible to carry out a transesterification and esterification reaction simultaneously in only

one stage, innovating over the conventional production process that uses two stages. In

addition, the pelleted biomass fly ash, with the characteristics of those used in this work,

allows to achieve high yields to FAME in both batch and continuous reactor. The production of

FAME in a continuous fixed-bed reactor using pelletized biomass fly ashes is a novelty of this

work.

In short, it is important to make the most of residual raw materials such as biomass fly ash or

WCO, as they represent a contribution on strengthening the circular economy by making good

use of these wastes for the energy generation (biodiesel). In this context, this research work

offers an approach that could allow minimizing the environmental impact and costs generated

by biodiesel production.

5.2 Future works

Based on the analysis carried out in the present study, some proposals for future studies are

suggested:

Complement characterization of the acidity and basicity on the surface of the evaluated

solid catalyst (biomass fly ash) using techniques such as temperature-programmed

desorption with NH3 and CO2. This characterization would assist the interpretation of

the results obtained regarding the catalytic performance of the material.

Carry out a characterization of the final products (FAME) obtained with the raw

materials and catalysts used in order to know their properties and compare them with

the specifications given by the ASTM D-6751 standard (or other).

FINAL REMARKS

127

It would be convenient to increase the operating time in the evaluation of the stability

of the pelletized catalyst in the continuous fixed-bed reactor, until a significant drop in

the catalytic activity is found.

Carry out a kinetic study of the solid biomass fly ash catalyst in the continuous fixed

bed reactor to find a representative kinetic model. Then, proceed to scale-up the

process, for example to a pilot scale, designing the continuous fixed-bed reactor to

perform assays with the solid pelleted fly ash catalyst to obtain data in conditions that

are more real and close to industrial ones.

Carry out a preliminary cost study for the production of FAME using several WCO

incorporation rates (i.e., several scenarios) and using the biomass fly ash catalyst. A life

cycle assessment study is also recommended in order to estimate the potential

environmental benefits of recovering these new raw (waste) materials in the FAME

production process.


128

Annexes

A1

Annexes

In the four articles published in co-authorship, the contribution of the author of this thesis was:

i. Conceptualization of the research work supported by an exhaustive and detailed

review of scientific literature;

ii. Design of research methodologies (both experimental and statistical) to achieve the

objectives set out;

iii. Execution of the following experimental activities : preparation of the catalysts, runs of

the chemical reactions in the laboratory reactors, gas chromatography for the

quantification of the products, design and construction of the continuous fixed-bed

reaction system, characterization of solid catalysts by FTIR and Hammett indicators,

physical-chemical characterization of oily raw materials;

iv. Validation and formal analysis of the experimental and statistical results obtained;

v. Both financial and laboratory resources (for de activities of previous point iii);

vi. Writing of article draft and review (in the process of submission to the scientific

journal).

edgar mauricio produÇÃo de biodiesel usando …

Documents